Releases

Project Briefs now recover automatically from a transient detail-shard read

A published airport could occasionally open the truthful unavailable state even though its reviewed detail record and evidence shard were healthy. Selected-record delivery now retries the static shard read at the edge and retries the brief or specialist request in the interface before showing an error.

Preserved: the airport record, source evidence, cost and schedule meanings, map position, spatial precision, discovery state, selected-marker framing, and explicit failure fallback are unchanged. A genuine repeated failure still produces the existing retryable unavailable state instead of inventing a brief.

Filters and shared-location cycling deepen discovery without changing project facts

The optional BEE layer adds operator and connector filters. Core discovery adds progress, compatible cost and deadline movement, exact-role families, and stable marker cycling when reviewed records share one source point.

Limit: missing or incomparable values remain separate. Role families are search aids; Project Briefs and Evidence retain exact source wording, without implying ownership, operation, blame, or misconduct.

Four hundred and eight cold-chain project entries expand logistics coverage without turning districts into facility sites

The catalog contains 11,027 source-backed records: 10,938 core non-toll records and 89 optional toll-plaza facilities. The separate optional EV charging lane remains 27,922 accepted facilities and contributes zero to the core inventory count.

Identity and role limit: four normalized names repeat in separately numbered approval rows with different approval dates. They remain distinct source-established entries without inferred phase or parent relationships. The source does not establish owner, operator, contractor, concessionaire, or EPC roles.

Local validation: source, money, spatial, containment, read-model, and D1 validators passed with clean idempotent replay and static parity. The 498-test suite, lint, minified production build, healthy and failed-basemap review, and responsive checks at 320, 390, 820, 1180, 1440, and 1920 pixels passed without horizontal overflow or browser errors.

Release integrity: the validated source, generated discovery artifacts, and ordered D1 migrations are versioned together. Production D1 is activated only after clean replay, semantic parity, and public-route verification.

Forty-one Mega Food Park projects add a complete Ministry inventory without flattening status, grants, or geography

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The candidate catalog contains 10,619 source-backed records: 10,530 core non-toll records and 89 optional toll-plaza facilities. Public production remains Release 97 with 10,204 records.

Role and evidence limit: each published SPV is labelled Source-listed SPV / implementing agency. It is not relabelled as owner, operator, contractor, concessionaire, or EPC provider, and procurement, payments, current lifecycle, and exact geometry remain missing unless separately sourced.

Account boundary: no Sites version, deployment, remote D1 write, hosting change, or remote push occurred. Publication remains queued behind an ordered migration audit, clean replay, complete release gate, and production verification from the original account.

Eighteen evidence-dated non-major-port projects expand the inventory without forcing unresolved geography

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The candidate catalog contains 10,578 source-backed records: 10,489 core non-toll records and 89 optional toll-plaza facilities. The separate 27,922-facility EV lane remains unchanged and does not count toward the core inventory. Public production remains Release 97 with 10,204 records.

Spatial and entity limit: all accepted markers are representative. Honnavar's two rows remain linked with their conflicting 4.9 versus 5 MTPA and ₹613 crore versus ₹607.03 crore observations visible in the ledger.

Account boundary: no Sites version, deployment, remote D1 write, hosting change, or remote push occurred. Publication remains queued behind an ordered migration audit, clean replay, complete release gate, and production verification from the original account.

The current official NIXI directory resolves all 79 facilities while preserving historic lifecycle boundaries

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The candidate catalog contains 10,560 source-backed records: 10,471 core non-toll records and 89 optional toll-plaza facilities. The separate 27,922-facility EV lane remains unchanged and does not count toward the core inventory. Public production remains Release 97 with 10,204 records.

Privacy and evidence limit: source-published addresses are retained, while contact names, phone numbers, and email addresses remain only in the frozen audit source and are omitted from product records. Exact exchange racks, buildings, routes, contracts, costs, and current performance remain missing.

Account boundary: no Sites version, deployment, remote D1 write, hosting change, or remote push occurred. Publication remains queued behind an ordered migration audit, clean replay, complete release gate, and production verification from the original account.

Forty-two historic NIXI Internet Exchange facilities expand Digital & Telecom without inventing the current location gap

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The candidate catalog contains 10,523 source-backed records: 10,434 core non-toll records and 89 optional toll-plaza facilities. The separate 27,922-facility EV lane remains unchanged and does not count toward the core inventory. Public production remains Release 97 with 10,204 records.

Spatial limit: Chennai 1 and Chennai 2 remain separate source records at the same representative Chennai anchor. Chandivali uses broader Mumbai orientation because no unambiguous reviewed locality point was available. No coordinate is jittered or upgraded to facility-site precision.

Account boundary: no Sites version, deployment, remote D1 write, hosting change, or remote push occurred. Publication remains queued behind an ordered migration audit, clean replay, complete release gate, and production verification from the original account.

Seventy deferred CEA rows are now auditable without creating State-centroid projects

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The candidate catalog remains 10,481 source-backed records: 10,392 core non-toll records and 89 optional toll-plaza facilities. The separate 27,922-facility EV lane remains unchanged and does not count toward the core inventory. Public production remains Release 97 with 10,204 records.

Evidence limit: needs-review is not rejection and does not prove that the underlying projects are invalid, inactive, or absent. A later authoritative project source may resolve identity and geography.

Account boundary: no Sites version, deployment, remote D1 write, hosting change, or remote push occurred. Publication remains queued behind an ordered migration audit, clean replay, complete release gate, and production verification from the original account.

Twelve locality-backed CEA renewable projects extend the inventory while sixteen uncertain rows remain held

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The candidate catalog contains 10,481 source-backed records: 10,392 core non-toll records and 89 optional toll-plaza facilities. The separate 27,922-facility EV lane remains unchanged and does not count toward the core inventory. Public production remains Release 97 with 10,204 records.

Evidence limit: the CEA table publishes no project coordinates or geometry. It reports developers, places, capacities, and schedule columns but does not establish current output, physical progress, cost, expenditure, contractor roles, or post-cutoff operation.

Account boundary: no Sites version, deployment, remote D1 write, hosting change, or remote push occurred. Publication remains queued behind an ordered migration audit, clean replay, complete release gate, and production verification from the original account.

Ten locality-backed CEA renewable projects extend the inventory while one blank-name State-only row remains held

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The candidate catalog contains 10,469 source-backed records: 10,380 core non-toll records and 89 optional toll-plaza facilities. The separate 27,922-facility EV lane remains unchanged and does not count toward the core inventory. Public production remains Release 97 with 10,204 records.

Evidence limit: the CEA table publishes no project coordinates or geometry. It reports developers, places, capacities, and schedule columns but does not establish current output, physical progress, cost, expenditure, contractor roles, or post-cutoff operation.

Account boundary: no Sites version, deployment, remote D1 write, hosting change, or remote push occurred. Publication remains queued behind an ordered migration audit, clean replay, complete release gate, and production verification from the original account.

Eighteen more CEA renewable projects extend the inventory while twenty-two ambiguous rows remain held

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The candidate catalog contains 10,459 source-backed records: 10,370 core non-toll records and 89 optional toll-plaza facilities. The separate 27,922-facility EV lane remains unchanged and does not count toward the core inventory. Public production remains Release 97 with 10,204 records.

Evidence limit: the CEA table publishes no project coordinates or geometry. It reports developers, places, capacities, and schedule columns but does not establish current output, physical progress, cost, expenditure, contractor roles, or post-cutoff operation.

Account boundary: no Sites version, deployment, remote D1 write, hosting change, or remote push occurred. Publication remains queued behind an ordered migration audit, clean replay, complete release gate, and production verification from the original account.

Fifteen more CEA renewable projects extend the inventory while eight uncertain source rows remain held

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The candidate catalog contains 10,441 source-backed records: 10,352 core non-toll records and 89 optional toll-plaza facilities. The separate 27,922-facility EV lane remains unchanged and does not count toward the core inventory. Public production remains Release 97 with 10,204 records.

Held evidence: two rows have zero commissioned and zero under-construction capacity, four publish Not Finalized as the project name, one combines Rajasthan solar and Karnataka wind without defensible point geometry, and one uses an organization name as both project and developer without a stable project identity.

Spatial limit: the Balai Kesharpura lookup returned no feature, so the Tejorupa record uses a broader Sheo locality point with the conflict stated directly. No accepted marker is CEA project geometry.

Account boundary: no Sites version, deployment, remote D1 write, hosting change, or remote push occurred. Publication remains queued behind an ordered migration audit, clean replay, complete release gate, and production verification from the original account.

Twelve more CEA renewable projects extend the reviewed inventory while conflicting and unconfirmed rows stay off the map

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The candidate catalog contains 10,426 source-backed records: 10,337 core non-toll records and 89 optional toll-plaza facilities. The separate 27,922-facility EV lane remains unchanged and does not count toward the core inventory. Public production remains Release 97 with 10,204 records.

Held evidence: one 640 MW row has an internally inconsistent zero commissioned and zero under-construction split; one Serentica row conflicts between Bhadla and Bikampur; and three Tilaiya or Panchet rows do not explicitly establish their apparent relationship to the existing DVC Floating Solar Park Phase I.

Account boundary: no Sites version, deployment, remote D1 write, hosting change, or remote push occurred. Publication remains queued behind an ordered migration audit, clean replay, complete release gate, and production verification from the original account.

Twelve more CEA renewable projects extend the reviewed execution inventory while nineteen uncertain or duplicate rows stay out of the count

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The candidate catalog contains 10,414 source-backed records: 10,325 core non-toll records and 89 optional toll-plaza facilities. The separate 27,922-facility EV lane remains unchanged and does not count toward the core inventory. Public production remains Release 97 with 10,204 records.

Source differences: Mirzapur and Odisha capacity observations differ from earlier canonical solar-park records, so they remain linked evidence conflicts rather than overwriting the existing values. Bagodara preserves the source difference between its name and the Surendranagar location column.

Local validation: source and spatial validators, clean idempotent D1 replay, D1/static parity, the complete automated suite, lint, the minified production build, responsive rendered review at 320, 390, 820, and 1440 pixels, selected-record state restoration, and deliberate basemap-failure isolation passed. No fresh 1180- or 1920-pixel browser capture was produced for this data-only checkpoint.

Account boundary: no Sites version, deployment, remote D1 write, hosting change, or remote push occurred. Publication remains queued behind an ordered migration audit, clean replay, complete release gate, and production verification from the original account.

Twenty-seven CEA renewable projects add source-scoped capacity and schedule evidence without duplicating two solar-park systems

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The candidate catalog contains 10,402 source-backed records: 10,313 core non-toll records and 89 optional toll-plaza facilities. The separate 27,922-facility EV lane remains unchanged and does not count toward the core inventory. Public production remains Release 97 with 10,204 records.

Evidence limit: CEA does not publish project coordinates or geometry in this table. The source role remains Source-listed developer, and a March 2026 status is not carried forward as current without newer evidence.

BESS source gap: the reviewed April 2026 CEA battery-storage report contains national aggregates but no project identities. The linked May 2026 PDF returned 404 during review, so neither source produced inventory records.

Account boundary: no Sites version, deployment, remote D1 write, hosting change, or remote push occurred. Publication remains queued behind an ordered migration audit, clean replay, complete release gate, and production verification from the original account.

Fifty-two CEA solar-park systems expand the energy inventory while three ambiguous locations remain held

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The Release 110 candidate contained 10,375 source-backed records: 10,286 core non-toll records and 89 optional toll-plaza facilities. The separate 27,922-facility EV lane remained outside the core inventory. Public production remained Release 97 with 10,204 records.

Evidence limit: the table can describe several nested plants, land parcels, transmission components, tenders, and commissioning events in one row. Those narratives remain attributable evidence and are not collapsed into one synthetic progress, cost, schedule, or lifecycle verdict.

Account boundary: no Sites version, deployment, remote D1 write, hosting change, or remote push occurred. Publication remains queued behind an ordered migration audit, clean replay, complete release gate, and production verification from the original account.

Twenty-one CEA hydro projects extend the execution inventory without duplicating nineteen existing records

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The candidate catalog contains 10,323 source-backed records: 10,234 core non-toll records and 89 optional toll-plaza facilities. The separate 27,922-facility EV lane remains unchanged and does not count toward the core inventory. Public production remains Release 97 with 10,204 records.

Evidence limit: the workbook does not publish reusable project coordinates or complete geometry. It also does not by itself establish original deadlines, revised deadlines, present progress, project cost, award, expenditure, payment, actual cost, final cost, or current operation.

Account boundary: no Sites version, deployment, remote D1 write, hosting change, or remote push occurred. Publication remains queued behind an ordered migration audit, clean replay, complete release gate, and production verification from the original account.

Fifteen operational land ports add a bounded cross-border logistics facility lane

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The candidate catalog contains 10,302 source-backed records: 10,213 core non-toll records and 89 optional toll-plaza facilities. The separate 27,922-facility EV lane remains unchanged and does not count toward the core inventory. Public production remains Release 97 with 10,204 records.

Source difference: the newer Ministry operational statement governs Sabroom's dated current read, while an older LPAI detail page retaining September 2021 pre-construction wording remains visible as a source difference rather than being erased.

Account boundary: no Sites version, deployment, remote D1 write, hosting change, or remote push occurred. Publication remains queued behind an ordered migration audit, clean replay, complete release gate, and production verification from the original account.

Seven reviewed Gandhinagar mine assets extend inspection evidence without turning locality landmarks into mine sites

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The candidate catalog contains 10,287 source-backed records: 10,198 core non-toll records and 89 optional toll-plaza facilities. The separate 27,922-facility EV lane remains unchanged and does not count toward the core inventory. Public production remains Release 97 with 10,204 records.

Privacy and source boundary: contact and inspection-administration fields in the public reports are excluded. Five register/report date conflicts, invalid or missing stable codes, unavailable reports, and unresolved geography remain held.

Account boundary: no Sites version, deployment, remote D1 write, hosting change, or remote push occurred. Publication remains queued behind an ordered migration audit, clean replay, complete release gate, and production verification from the original account.

Twenty-four reviewed mine assets add current IBM inspection evidence without inventing operation or mine-site precision

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The candidate catalog contains 10,280 source-backed records: 10,191 core non-toll records and 89 optional toll-plaza facilities. The separate 27,922-facility EV lane remains unchanged and does not count toward the core inventory. Public production remains Release 97 with 10,204 records.

Privacy and source boundary: the public reports contain contact and inspection-administration fields. The product extracts only non-sensitive project identity, exact source role, dated inspection, mineral, administrative place, and reviewed mine attributes needed for infrastructure discovery.

Account boundary: no Sites version, deployment, remote D1 write, hosting change, or remote push occurred. Publication remains queued behind an ordered migration audit, clean replay, complete release gate, and production verification from the original account.

Seven reviewed public space-launch records activate a dedicated infrastructure category

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The candidate catalog contains 10,256 source-backed records: 10,167 core non-toll records and 89 optional toll-plaza facilities. The separate 27,922-facility EV lane remains unchanged and does not count toward the core inventory. Public production remains Release 97 with 10,204 records.

Relationship and source boundary: four source-backed child relationships connect the First, Second, and Third Launch Pads and Agnikul Launchpad-01 to SDSC SHAR. Shared Sriharikota anchors are preserved without coordinate jitter or record merging.

Held candidates: Skyroot's Infinity Campus and Max-Q Campus remain needs review because the official sources reviewed here do not yet reconcile their names, sites, and relationship strongly enough for canonical publication.

Account boundary: no Sites version, deployment, remote D1 write, hosting change, or remote push occurred. Publication remains queued behind an ordered migration audit, clean replay, complete release gate, and production verification from the original account.

Seven reviewed shipbuilding and ship-repair records activate a dedicated infrastructure category

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The candidate catalog contains 10,249 source-backed records: 10,160 core non-toll records and 89 optional toll-plaza facilities. The separate 27,922-facility EV lane remains unchanged and does not count toward the core inventory. Public production remains Release 97 with 10,204 records.

Money and source boundary: ₹1,799 crore, ₹970 crore, ₹168 crore, ₹180 crore, ₹175.20 crore, and ₹70 crore observations retain their published dates and definitions. No compatible cost-revision sequence is inferred.

Held proposals: the official 2025 Kochi block-fabrication and Tamil Nadu greenfield-shipyard proposals remain needs review because the sources do not establish a publishable approved site and source-backed map position. The reported Tata Kerala facility remains outside the ledger because no authoritative reusable source was located.

Account boundary: no Sites version, deployment, remote D1 write, hosting change, or remote push occurred. Publication remains queued behind an ordered migration audit, clean replay, complete release gate, and production verification from the original account.

EV charging discovery moves from a full browser catalog to bounded D1-first reads

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The core app catalog remains 10,242 records and the 27,922 accepted EV facilities contribute zero to the core 10,000-project count. Public production remains Release 97 with 10,204 records.

Scale boundary: a dense Maharashtra close-view response measured roughly 507 KB for 1,850 active points. The server fallback also parses and caches the full static source on its first request. Production D1 latency and cache behavior remain unverified until the original owning account returns.

Account boundary: no Sites version, deployment, remote D1 write, hosting change, or remote push occurred. Publication remains queued behind an ordered migration audit, clean replay, complete release gate, and production verification from the original account.

Reviewed EV charging facilities become an explicit, progressively loaded discovery layer

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The core app catalog remains 10,242 records and the 27,922 accepted EV facilities contribute zero to the core 10,000-project count. Public production remains Release 97 with 10,204 records.

Scale boundary: the compressed static index is an independently healthy local fallback and proves the interaction contract. Production-scale viewport queries and cursor pagination should move to D1 after the original owning account returns and exact parity is revalidated.

Account boundary: no Sites version, deployment, remote D1 write, hosting change, or remote push occurred. Publication remains queued behind an ordered clean-database replay and complete release gate from the original account.

Nationwide BEE public EV charging inventory staged without claiming present operation

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. This checkpoint does not change the active 10,242-record app catalog or the core 10,000-project count. Public production remains Release 97 with 10,204 records.

D1 boundary: eight local SQL artifacts preserve all accepted and held dispositions in a dedicated candidate table. Clean SQLite replay and a second application retain identical candidate and search counts. No migration has been applied remotely.

Interface boundary: the active spatial index remains 10,242 records and contains no EV charging facilities. Search, map, URL, result, selection, and optional-layer integration remain a later validated checkpoint.

Account boundary: publication remains queued until the original Sites-owning account returns, accumulated migrations replay in order, and the exact candidate passes the full release gate again.

Sleemanabad and Bargi Diversion Phase III linked without creating a duplicate project

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. This candidate is not saved, deployed, or applied to remote D1. The candidate catalog remains 10,242 source-backed records: 10,153 core non-toll records and 89 optional toll-plaza facilities. Public production remains Release 97 with 10,204 records.

Money boundary: ₹799 crore turnkey award, ₹9.76 crore grout-block contract, ₹1,356.05 crore running-bill work value, and the ₹13.24 crore, ₹39.60 crore, and ₹6.20 crore audit observations retain different meanings and are not compared as a cost-revision sequence.

Status boundary: the parent's June 2026 snapshot remains 81.52% progress with a December 2026 revised target. Historical excavation, monitoring, extension, and under-execution evidence does not establish tunnel breakthrough, completion, commissioning, water release, or operation.

Account boundary: publication remains queued until the original Sites-owning account returns, accumulated migrations replay in order, and the exact candidate passes the full release gate again.

Bhogapuram airport development added without collapsing construction, licensing, opening, or cost meanings

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The candidate catalog contains 10,242 source-backed records: 10,153 core non-toll records and 89 optional toll-plaza facilities. Public production remains Release 97 with 10,204 records.

Source boundary: four official HTML sources are frozen and fingerprinted. The public AAI coordinate PDF was reviewed, but repeated direct downloads timed out, so the linked coordinate remains attributable without a claimed immutable local PDF hash.

Containment boundary: the official point passes coordinate-order, India-bounds, Andhra Pradesh, Vizianagaram, airport-identity, and source-context review. The frozen Release 97 polygon snapshots remain unchanged, so no newer independent polygon-containment result is claimed.

Account boundary: publication remains queued until the original Sites-owning account returns, accumulated migrations replay in order, and the exact candidate passes the full release gate again.

Thirty-seven PPPAC Port projects prepared locally with appraisal and revision meanings intact

LOCAL ONLY, NOT SAVED, NOT DEPLOYED, REMOTE D1 UNCHANGED. The candidate catalog contains 10,241 source-backed records: 10,152 core non-toll records and 89 optional toll-plaza facilities. Public production remains Release 97 with 10,204 records.

Status boundary: PPPAC final or in-principle approval is dated appraisal evidence. It does not establish current procurement, concession, construction, completion, commissioning, opening, operation, or ownership.

Containment boundary: the frozen Release 97 state and district snapshots still cover the 10,204 released records. The 37 local additions pass frozen-source coordinate, State/UT, India-bounds, precision, and limitation checks, but no new independent polygon-containment result is claimed.

Account boundary: publication remains queued until the original Sites-owning account returns, the accumulated local work is audited, migrations are replayed in order, and the exact candidate passes the full release gate again.

Ninety-nine Smart Cities ICCC systems with dated status and locality-level maps

The public catalog expands to 10,194 source-backed records: 10,105 core non-toll records and 89 optional toll-plaza facilities. Digital & Telecom grows from 37 to 136 records through a nationwide Integrated Command and Control Centre baseline lane.

Identity boundary: the OGD city list establishes the 100-city Smart Cities Mission universe and selection round under the Government Open Data License - India. It does not establish a control-room site, contractor, cost, deadline, capacity, or current status.

Status boundary: the 2024 nationwide MoHUA statement does not prove uninterrupted service, current integrations, performance, capacity, ownership, commissioning scope, or present operation in any city.

Coverage boundary: city-specific tenders, contracts, amendments, costs, payments, audits, operators, stronger geometry, and newer status evidence remain explicit enrichment gaps.

Zojila Tunnel contract trail with its source conflict intact

The public catalog expands to 10,095 source-backed records: 10,006 core non-toll records and 89 optional toll-plaza facilities. The held PAIMANA candidate is now resolved through dedicated NHIDCL and PIB evidence without rewriting its incorrect Andhra Pradesh state field.

Schedule boundary: the current NHIDCL row moves the likely completion from 30 September 2026 to 28 February 2028, a 516-day change. Its 69% physical and 63% financial figures use the 19 July retrieval date because the row publishes no separate observation date.

Source-difference boundary: PAIMANA's Andhra Pradesh field, the 2024 PIB length/cost/schedule figures, and the rounded 2026 breakthrough values remain visible differences rather than silent corrections.

Evidence boundary: no signed EPC agreement, amendments, certified payment ledger, disputes, audits, actual cost, final cost, completion, commissioning, opening, or operation is inferred.

Vadhavan Port as a distinct contract-linked development project

The public catalog expands to 10,094 source-backed records: 10,005 core non-toll records and 89 optional toll-plaza facilities. Vadhavan is now represented as a development project, not an operating port asset or an alias for its separate highway-connectivity packages.

Cost boundary: the package estimate and package award are compatible with each other but are not whole-project original or revised costs. No actual, final, expenditure, payment, or whole-project award value is inferred.

Lifecycle boundary: Cabinet approval and a dated package award establish the reported project stage through 31 December 2024. They do not establish physical progress, completion, commissioning, inauguration, opening, or current operation.

Coverage boundary: this resolves the held Vadhavan development-project candidate from the major-port lane. Galathea Bay remains held, and the separately modelled NHAI connectivity records remain distinct.

A compact default discovery response without losing individual projects

The default 10,004-record map no longer receives a redundant list of all 10,004 record IDs from D1. The already loaded spatial index applies the truthful non-toll catalog scope locally, while filtered, searched, and awareness states continue receiving exact explicit identities.

Serving boundary: compact activation is allowed only when the discovery layer is All records and no filter or search changes membership. Toll opt-in selects the complete reviewed catalog scope. Any narrower state falls back to explicit IDs.

Truth boundary: the catalog, category, lifecycle, evidence, money, schedule, organization, coordinate, precision, marker, icon, selected dossier, and default toll meanings do not change.

Remaining scale boundary: broad filtered searches can still return large explicit-ID sets. Later viewport activation must preserve exact counts, accessible directory access, source coordinates, and every individual unclustered record.

Production D1 discovery with automatic static fallback

The public catalog remains 10,093 source-backed records. Filtered browse, facets, awareness questions, toll opt-in, and cursor pages now use a version-gated D1 read model when healthy, without changing the map, evidence, taxonomy, icons, or selected-project dossiers.

Serving boundary: D1 becomes primary only when its discovery model reports the exact public data version. Static search, results, selected-record detail shards, evidence, and map-failure access remain independently available.

Scale boundary: this release removes the routine 23.9 MB server-side current-facts fetch from healthy filtered discovery. It still returns the complete active ID set for individual unclustered map activation; viewport query and full-text search remain later scale work.

Truth boundary: no project, coordinate, category, lifecycle, money, schedule, organization role, evidence field, toll default, icon, or dossier content changes in this release.

Ten thousand core infrastructure records

The public catalog expands to 10,093 source-backed records and crosses the inventory milestone with 10,004 core non-toll projects, packages, assets, and source-defined systems.

Truth boundary: official Ministry of Railways and NHAI records establish identity and dated source status. GeoNames and OpenStreetMap provide representative navigation only and do not establish project geometry, present operation, award, contractor, cost, schedule, or progress.

Relationship boundary: four source-named NHAI packages link to a broader parent scope only where the published names and chainages support the relationship. Unresolved phase, overlap, and duplicate questions remain held.

D1 and serving boundary: two idempotent migration shards preserve the 438-row control plane, two duplicate decisions, four relationships, and 260 baseline records. Ordered-query parity matches the static fallback across catalog, default scope, state, category, lifecycle, evidence, awareness, and search queries.

Next phase: the numeric inventory milestone is a foundation, not a completeness claim. Source refresh, underrepresented categories, stronger geometry, responsible roles, contracts, money, schedules, audits, and deeper evidence continue.

One thousand six hundred eighty-one AMRUT projects with exact-name locality navigation

The public catalog expands to 9,833 source-backed records without redesigning the accepted map, freestanding pictograms, discovery workspace, compact Project Brief, optional toll layer, or specialist evidence model.

Truth boundary: portal Project Cost remains its own source-defined quantity, and an undated milestone is not carried forward as current construction, completion, commissioning, inauguration, opening, or operation.

Spatial boundary: acceptance requires one exact normalized primary-or-alternate-name GeoNames place feature in the official source state. A locality point is not a project site, facility, service area, network, route, parcel, or surveyed extent. Barrackpore and North Barrackpore share one source feature without merging their official identities or moving the point.

D1 and serving boundary: eight idempotent migration shards add the 1,681 baseline records, preserve all residual dispositions, and normalize separate official AMRUT and GeoNames source links. Static browse, selected-record evidence, and map-failure access remain independent fallbacks.

Remaining gap: 9,744 core records represent 97.44% of the 10,000-record target. Another 256 accepted core projects or infrastructure assets are required. The held or unsupported AMRUT rows will not be used merely to close that gap.

Thirty-eight historical CEA generation assets with conservative representative navigation

The public catalog expands to 8,152 source-backed records without redesigning the accepted map, freestanding pictograms, discovery workspace, compact Project Brief, optional toll layer, or specialist evidence model.

Truth boundary: CEA establishes the station identity and January 2024 observation. It publishes no coordinate in this table. The map point comes from a separately attributed ODbL OpenStreetMap feature and remains representative rather than official or exact.

Selection boundary: acceptance requires a unique named power-plant feature in the independently verified source state, compatible fuel, a normalized match score of at least 0.70, a margin of at least 0.15 over the next candidate, and no high-confidence current-catalog match. The other 72 candidates remain held rather than receiving weak or invented geography.

D1 and serving boundary: one idempotent migration shard adds 38 baseline records and all 110 candidate dispositions. Static browse, selected-record evidence, and map-failure access remain independent fallbacks.

Remaining gap: 8,063 core records represent 80.63% of the 10,000-record target. Another 1,937 accepted core projects or infrastructure assets are required.

Two hundred more reviewed AMRUT 2.0 projects from proven locality anchors

The public catalog expands to 8,114 source-backed records without redesigning the accepted map, freestanding pictograms, discovery workspace, compact Project Brief, optional toll layer, or specialist evidence model.

Truth boundary: Release 88 reuses the frozen AMRUT 2.0 directory and reviewed source semantics from Release 79. It does not turn portal Project Cost into original, sanctioned, revised, awarded, paid, actual, or final cost, and it does not carry an undated milestone forward as current construction, completion, commissioning, or operation.

Spatial boundary: the Ministry directory publishes no project coordinate or geometry. Each record uses a separately attributed locality representative already reviewed for its ULB and state. Shared coordinates remain explicit; no site, treatment plant, pipe, sewer, service area, parcel, or route is invented.

D1 and serving boundary: three idempotent migration shards add 200 baseline records and preserve all 7,612 residual candidate dispositions. The compact spatial index, cursor discovery path, selected-record detail shards, and static failure fallback remain independently usable.

Rendered review: the exact production build was reviewed at 320, 390, 820 portrait, 1180 landscape, 1280, 1440, and 1920 pixels without page-level horizontal overflow or browser console errors. The 8,025-record default non-toll scope, 8,114-record catalog, Pandameru-3 search and selection, representative-locality wording, ₹4.74 crore portal Project Cost boundary, Back and Home continuity, phone detail sheet, tablet selected-marker framing, pan, zoom, India reset, and deliberately blocked-basemap directory remained usable.

Remaining gap: 8,025 core records represent 80.25% of the 10,000-record target. Another 1,975 accepted core projects or infrastructure assets are required. This reviewed-anchor batch is not a complete current national water or sewerage inventory.

Three hundred five civic-utility facilities from Smart Cities open data

The public catalog expands to 7,914 source-backed records without redesigning the accepted map, freestanding pictograms, discovery workspace, compact Project Brief, optional toll layer, or specialist evidence model.

Truth boundary: the source resources are city asset inventories rather than current operational registers. Release 87 reports Listed in dated city inventory or Commissioning year published; it does not convert a historical listing or year into present operation, continuous service, ownership, capacity, progress, completion, or performance.

Spatial boundary: accepted coordinates are published facility points displayed in WGS 84 longitude-latitude order. Every accepted point matches its declared state and one district, but the point is not upgraded into a surveyed parcel, building footprint, service area, pipe, cable, network, or municipal boundary. Three independently detected state conflicts remain needs review without moving coordinates.

D1 and serving boundary: one idempotent migration shard adds 305 baseline records, all 448 candidate dispositions, five explicit duplicate decisions, 305 typed source-row observations, exact resource links, and official facility-point spatial records. Static browse and evidence reads remain the independent fallback.

Rendered review: the exact production build was reviewed at 320, 390, 820 portrait, 1180 landscape, 1280, 1440, and 1920 pixels without page-level horizontal overflow or browser console errors. The 7,825-record default non-toll scope, 7,914-record catalog, Sanjay Place CWR search and selection, official-facility precision wording, category filters, toll opt-in, Back and Home continuity, phone detail sheet, tablet workspace, selected-marker framing, zoom and Reset to India controls, and deliberately blocked-basemap directory remained usable.

Remaining gap: 7,825 core records represent 78.25% of the 10,000-record target. Another 2,175 accepted core projects or infrastructure assets are required. This bounded nine-state civic-utility source lane is not a complete national utility inventory.

Thirty-four BharatNet State/UT systems with source-dated service-ready counts

The public catalog expands to 7,609 source-backed records without redesigning the accepted map, freestanding pictograms, discovery workspace, compact Project Brief, optional toll layer, or specialist evidence model.

Truth boundary: service-ready points are not unique GPs, routes, towers, exchanges, contracts, completion percentages, commissioned networks, operating connections, or performance verdicts. State-specific money, contractors, present operation, and completion remain not published.

Spatial boundary: the annexure publishes no physical coordinate or network geometry. Survey of India cartographic State/UT label placements are administrative navigation anchors only. Four sit outside the independent land polygons and Puducherry intersects Tamil Nadu; all discrepancies remain explicit without moving coordinates.

D1 and serving boundary: one idempotent migration shard adds 34 baseline records, all 34 candidate dispositions, shared programme relationships, typed count observations, 68 exact organization-role rows, official source links, and administrative spatial records. Static browse and evidence reads remain the independent fallback.

Rendered review: the exact production build was checked at 320, 390, 820 portrait, 1,180 landscape, 1,280, 1,440, and 1,920 pixels. BharatNet search returned all 34 State/UT systems, the compact Project Brief kept its administrative precision and source gaps, Back restored the search, Home restored the clean 7,520-record non-toll scope, and a blocked basemap still left search and all 24 first-page records usable without horizontal overflow or browser errors.

Remaining gap: 7,520 core records represent 75.2% of the 10,000-record target. Another 2,480 accepted core projects are required. This source establishes State/UT system scopes, not a physical inventory of every BharatNet facility or route.

Thirty-six Namami Gange sewerage projects with historical status and separated money meanings

The public catalog expands to 7,575 source-backed records without redesigning the accepted map, freestanding pictograms, discovery workspace, compact Project Brief, optional toll layer, or specialist evidence model.

Truth boundary: Completed, Completed under O&M, Under Progress, tendering, and AA&ES wording are dated source observations and are not carried forward as current 2026 status. Approved project cost is not revised or final cost, and cumulative expenditure is not treated as a compatible cost revision or contractor payment.

Spatial boundary: the annexure publishes no project coordinate or geometry. Each accepted record uses a separately attributed OpenStreetMap locality representative, never an STP site, sewer or drain alignment, footprint, parcel, service area, or surveyed project extent.

D1 and serving boundary: one idempotent migration shard adds 36 baseline records, all 41 dispositions, two deduplication decisions, exact source links, typed status and row-field observations, source-reporting-ministry roles, and representative spatial records. Static browse and evidence reads remain the independent fallback.

Rendered review: the exact production build was reviewed at 320 and 390 phone widths, 820 tablet portrait, 1180 landscape, 1440, and 1920 pixels without page-level horizontal overflow or browser console errors. It verified the 7,486-record default non-toll scope, 436-record Environmental Utilities facet, Joshimath search and selection, compact Project Brief, approved cost and cumulative expenditure as separately typed money, representative-locality wording, selected-marker framing, Home reset, pan, zoom, India reset, phone and tablet detail, and a deliberately blocked-basemap directory.

Remaining gap: 7,486 core records represent 74.86% of the 10,000-record target. Another 2,514 accepted core projects are required. This five-state historical annexure is a bounded source lane, not a current national wastewater inventory.

One hundred sixteen major railway projects from a Ministry of Railways/PIB source lane

The public catalog expands to 7,539 source-backed records without redesigning the accepted map, freestanding pictograms, discovery workspace, compact Project Brief, optional toll layer, or specialist evidence model.

Spatial boundary: the PIB tables publish no reusable coordinate or railway alignment. Each accepted record uses a separately attributed OpenStreetMap named endpoint or locality whose own address passed India and source-state review. No route, midpoint, worksite, parcel, project boundary, or surveyed extent is invented.

D1 and serving boundary: one idempotent migration shard adds 116 baseline records, all 163 dispositions, 33 deduplication decisions, exact source links, typed status and row-field observations, source-reporting-ministry roles, and representative spatial records. Static browse and evidence reads remain the independent fallback.

Rendered review: the exact production build was reviewed at 320, 390, 820 portrait, 1024, 1180 landscape, 1280, 1440, and 1920 pixels without page-level horizontal overflow or browser console errors. Railway search, the 1,448-record Railway facet, Araria-Galgalia selection, compact Project Brief, literal project-cost wording, official evidence access, selected-marker framing, phone sheet, tablet workspace, default non-toll scope, and map-independent directory remained usable. A local wide-screen run painted the first real project markers at 590 ms and reported LCP at 296 ms; these are release evidence, not a public p75 claim.

Remaining gap: 7,450 core records represent 74.5% of the 10,000-record target. Another 2,550 accepted core projects are required. The four-state parliamentary answer is a bounded source lane, not an all-India railway inventory.

One hundred eighty-one NHAI national-highway projects and packages

The public catalog expands to 7,423 source-backed records without redesigning the accepted map, freestanding pictograms, discovery workspace, compact Project Brief, optional toll layer, or specialist evidence model.

Spatial boundary: the NHAI report publishes no reusable coordinate or route geometry. Each accepted record uses a separately attributed OpenStreetMap place or administrative representative. No alignment, chainage, package midpoint, worksite, land boundary, or surveyed extent is invented.

D1 and serving boundary: one idempotent migration shard adds 181 baseline records, all 821 dispositions, nine deduplication decisions, exact source links, field observations, source-publisher roles, and representative spatial records. Fourteen ordered query sets retain exact D1-to-static parity while static browse and evidence reads remain the independent fallback.

Rendered review: the exact production build was reviewed at 320, 390, 820 portrait, 1180 landscape, 1440, and 1920 pixels without page-level horizontal overflow or browser console errors. The 7,334-record default non-toll scope, 719-record Highway / Road facet, NHAI search and selection, compact Project Brief, representative-point wording, shareable filter URL, clean Home reset, and map-independent directory path remained usable while the accepted freestanding pictograms, selected-marker framing, and existing workspace behavior stayed unchanged.

Remaining gap: 7,334 core records represent 73.34% of the 10,000-record target. Another 2,666 accepted core projects are required, with category and state balance still governed by the inventory coverage matrix.

Seventeen unique NICDC industrial-corridor projects after canonical deduplication

The public catalog expands to 7,242 source-backed records without redesigning the accepted map, freestanding pictograms, discovery workspace, compact Project Brief, optional toll layer, or specialist evidence model.

Spatial boundary: NICDC publishes no reusable project coordinates or geometry in the reviewed report. Sixteen projects use separately attributed locality representative points. The Jharkhand record remains unresolved at state level because the report says an alternate site is being identified. None is presented as a project site, land boundary, route, or surveyed extent.

D1 and serving boundary: one idempotent migration shard adds 17 baseline records, all 27 candidate dispositions, ten source-to-canonical links, exact organization roles, typed stage and area observations, and spatial-source links. Thirteen ordered query sets retain exact D1-to-static parity while static browse and evidence reads remain the independent fallback.

Rendered review: the exact production build passed the selected-project flow at 320, 390, 820 portrait, 1180 landscape, 1280, 1440, and 1920 pixel widths without page-level horizontal overflow or a workspace-hidden marker. The first real project markers appeared in about 451 ms in the local 1440-pixel review; this is build evidence, not a production percentile or a claim that every device meets the long-term performance budget.

Remaining gap: 7,153 core records represent 71.53% of the 10,000-record target. Another 2,847 accepted core projects are required, with category and state balance still governed by the inventory coverage matrix.

Fifteen prioritised multimodal logistics parks

The public catalog expands to 7,225 source-backed records without redesigning the accepted map, freestanding pictograms, discovery workspace, compact Project Brief, optional toll layer, or specialist evidence model.

Truth boundary: the Ministry's 1 April 2026 Parliament answer supports named MMLP identities, locations, a dated implementation or DPR state, and authority-estimated cargo capacity for four rows. It does not establish present-day status, exact sites, cost, contractor, concessionaire, physical progress, completion, operation, or performance.

Spatial boundary: the answer publishes no coordinate or geometry. Each record uses a separately attributed OpenStreetMap city or locality representative point, never the MMLP site, facility extent, transport connection, land parcel, or surveyed position.

D1 and serving boundary: one idempotent migration shard adds 15 baseline records, all candidate dispositions, exact reporting-ministry roles, dated status observations, source links, and representative spatial records. Static browse and evidence reads remain the independent public fallback.

Remaining gap: 7,136 core records represent 71.36% of the 10,000-record target. Another 2,864 accepted core projects are required.

Fifty-six reviewed PNGRB pipeline assets

The public catalog expanded to 7,210 source-backed records without redesigning the accepted map, freestanding pictograms, discovery workspace, compact Project Brief, optional toll layer, or specialist evidence model.

Truth boundary: PNGRB's March 2026 physical-state tables and April 2026 monthly observations are dated evidence, not present-day status. The source column named Entity remains PNGRB-reported entity, not contractor, operator, owner, sponsor, concessionaire, or authority. Representative endpoint and administrative anchors are not pipeline routes, stations, rights of way, or surveyed geometry.

D1 and serving boundary: one idempotent migration shard adds 56 baseline records, all 65 candidate dispositions, nine deduplication decisions, exact source roles, typed physical and monthly observations, source links, and representative spatial records. Thirteen ordered query sets retain exact D1-to-static parity, while static browse and evidence reads remain the independent public fallback.

Rendered review: the exact production build was reviewed at 320, 390, 820 portrait, 1440, and 1920 pixels. The 7,121-record default non-toll directory, 126-record Pipelines & Fuel facet, a representative PNGRB Project Brief, exact organization-role wording, Back-state continuity, clean Home reset, source-locked freestanding icons, and deliberately blocked-basemap fallback remained usable without page-level horizontal overflow or browser warnings. A combined drag and wheel sample held 60.1 frames per second.

Remaining gap: 7,121 core records represented 71.21% of the 10,000-record target. Another 2,879 accepted core projects were required, with category and state balance still governed by the inventory coverage matrix.

Five hundred state-balanced AMRUT 2.0 water and sanitation records

The public catalog expands to 7,154 source-backed records without redesigning the accepted map, freestanding pictograms, discovery workspace, compact Project Brief, optional toll layer, or specialist evidence model.

Spatial boundary: the Ministry pages name ULBs but publish no project coordinate or geometry. The 500 projects use 374 separately attributed representative locality anchors. Shared anchors are preserved without jitter and do not imply project-site precision.

Review boundary: three unresolved ULB navigation gaps remain needs-review rather than receiving guessed coordinates. Smart Cities and PMGSY source paths did not meet this release's reproducible public-access contract and contributed no records.

D1 and serving boundary: four idempotent migration shards retain the exact 500 project identities and the complete 8,769-row candidate ledger. Twelve ordered query sets preserve D1-to-static parity while static browse and evidence reads remain the independent fallback.

Remaining gap: 7,065 core records represent 70.65% of the 10,000-record target. The next lane should improve transport, logistics, industrial, digital, pipeline, or environmental category balance rather than repeatedly selecting from the same urban-water source.

Forty-seven under-construction TBCB transmission projects after cross-catalog deduplication

An eleventh auditable source lane expands the public catalog to 6,654 source-backed records without redesigning the accepted map, freestanding pictograms, discovery workspace, compact Project Brief, optional toll layer, or specialist evidence model.

Source boundary: the CEA May 2026 TBCB report is indicative and states that it must not be used as a ground to claim time or cost overrun. Its 301 component rows have different physical denominators and are not averaged into one whole-project progress percentage.

Spatial boundary: the report publishes no reusable coordinate or line geometry. Forty-seven projects use 41 separately attributed representative OpenStreetMap anchors from source-named places. These are not CEA coordinates, substations, towers, rights-of-way, surveyed alignments, or complete project extents.

D1 and serving boundary: one idempotent migration shard adds 47 accepted records, the complete 86-row disposition ledger, 39 deduplication decisions, typed cost and schedule observations, exact Executing Agency roles, source-dated lifecycle events, source links, and representative spatial records. Twelve ordered query sets retain exact D1-to-static parity, while static browse and evidence reads remain the independent public fallback.

Rendered review: the exact production build was reviewed at 320, 390, 820 portrait, 1280, and 1920 pixels. The 6,565-record default non-toll directory, a distinctive new TBCB search, representative selected-marker framing, compact Project Brief, typed Money view, Back-state continuity, clean Home reset, optional toll layer, and deliberately blocked-basemap fallback remained usable without page-level horizontal overflow or browser warnings.

Remaining gap: 6,565 core records represent 65.65% of the 10,000-record target. Another 3,435 accepted core projects are required, with category and state balance still governed by the inventory coverage matrix.

Fifty-eight under-construction RTM transmission projects with source-dated lifecycle evidence

A tenth auditable source lane expanded the public catalog to 6,607 source-backed records without redesigning the accepted map, freestanding pictograms, discovery workspace, compact Project Brief, optional toll layer, or specialist evidence model.

Source boundary: the CEA report is indicative and states that it must not be used as a ground to claim time or cost overrun. Component indicators with different denominators are not averaged into synthetic whole-project progress.

Spatial boundary: 58 projects use 47 separately attributed representative anchors, not CEA coordinates, exact substations, routes, towers, or project extents.

Seventy-eight completed-TBCB transmission projects with typed schedule evidence

A ninth auditable source lane expands the public catalog to 6,549 source-backed records without redesigning the accepted map, pictograms, discovery workspace, compact Project Brief, optional toll layer, or specialist evidence model.

Source boundary: the CEA May 2026 completed-TBCB report is indicative and states that it must not be used as a ground to claim time or cost overrun. Completed and Actual COD do not establish present operation, availability, ownership, asset condition, utilization, or benefit realization.

Spatial boundary: the report publishes no reusable coordinate or line geometry. Seventy-eight projects use 57 separately attributed representative OpenStreetMap anchors from source-named places. These are not CEA coordinates, substations, towers, rights-of-way, surveyed alignments, or complete project extents.

D1 and serving boundary: one idempotent migration shard adds 78 accepted records, the complete 79-row disposition ledger, typed cost and schedule observations, exact Executing Agency roles, lifecycle events, source links, and representative spatial records. Twelve ordered query sets retain exact D1-to-static parity, while static browse and evidence reads remain the independent public fallback.

Rendered review: the exact production build was reviewed at 320, 390, 820 portrait, 1180 landscape, 1440, and 1920 pixels. The 6,460-record default non-toll scope, 131-record Power Grid & Storage facet, CEA Project Brief, typed Timeline and Evidence views, clean Home reset, selected-marker framing, and deliberate basemap-failure directory remained usable without page-level horizontal overflow or browser console errors.

Remaining gap: 6,460 core records represent 64.6% of the 10,000-record target. The separately staged CEA under-construction RTM and TBCB reports are not part of this release and require their own identity, spatial, lifecycle, schedule, and disclaimer review.

Freestanding category pictograms replace circular marker medallions

The map now gives every infrastructure category a fuller, clearer icon-first identity without changing any record, category assignment, coordinate, evidence field, filter, lifecycle, URL, dossier, toll default, or D1 inventory rule.

Subtype boundary: irrigation, renewable generation, data-centre, telecom or fibre, and submarine-cable variants are presentation-only within their canonical category. A variant activates only from a reviewed source-type or sector field, never from cost or a vague name. Unsupported variants remain dormant.

Performance boundary: markers remain native MapLibre symbols with feature-state interaction, source-locked coordinates, larger invisible hit targets, deterministic draw priority, no DOM markers, no clusters, no jitter, and no positional transition. The small real-point fallback remains available if an icon asset fails.

Visual boundary: the supplied full-resolution concept is art direction only. The shipped assets are generated from the maintained professional vector family rather than cropped from the poster, and the poster contributes no project data or taxonomy fact.

1,095 railway-station development records from a complete national scheme list

An eighth auditable source lane expands the public catalog to 6,471 source-backed records without redesigning the accepted map, discovery workspace, Project Brief, optional toll layer, or specialist evidence model.

Source boundary: the Ministry of Railways list published through PIB supplies station identity, state, scheme inclusion, and the source total. A later PIB list supplies the dated completion observation for matched station names. Neither source supplies station-level cost, contractor, tender, schedule, exact works, present operation, or universal current progress.

Spatial boundary: frozen Wikidata CC0 railway-station points are used only after exact normalized label or alias matching, India bounds, unique-item review, and independent state containment. A representative point is not an official Railway coordinate, station boundary, platform, parcel, construction worksite, or project footprint.

Lifecycle boundary: identified, completed, commissioned, inaugurated, opened, and presently operating remain different states. Completion is shown only for the 141 exact reviewed crosswalks; the absence of a station from the later 180-name completion list is not treated as proof that it remains incomplete.

D1 and serving boundary: five idempotent migration shards add 1,095 accepted records, the complete 1,309-position candidate ledger, typed lifecycle observations, exact scheme-ministry roles, navigation evidence, aliases, maturity rows, and source links. Static browse and evidence reads remain the independent public fallback.

Scale correction: the first saved deployment candidate failed safely before replacing production because the 50,513,456-byte worker exceeded the platform startup CPU budget. Full selected-record data now lives in 256 deterministic, on-demand static shards. The exact replacement worker is 734,512 bytes, while all 6,471 records, Project Briefs, specialist sections, evidence, and fallback paths remain intact.

Rendered review: the exact production build was reviewed at 320, 390, 820 portrait, 1180 landscape, 1440, and 1920 pixels. The 6,382-record default non-toll scope, unchanged Major-project layer, Facilities, Railway filtering, station search, Project Briefs, clean Home reset, phone detail sheet, tablet selected-marker framing, and static map-failure access remained usable without page-level horizontal overflow or browser console errors.

Remaining gap: 6,382 core records represent 63.82% of the 10,000-record target. Seventy-one ABSS candidates still need adjudication and 143 need stronger reusable navigation evidence. Large permission-cleared road, rail-project, logistics, industrial, grid, digital, urban, and environmental inventories remain the broader coverage bottleneck.

Sixty-eight National Waterway systems with strict representative geography

A seventh auditable source lane expands the public catalog to 5,376 source-backed records without redesigning the accepted map, discovery workspace, Project Brief, optional toll layer, or specialist evidence model.

Source boundary: the Ministry's National Waterways list supplies the stable identity, component descriptions, published component lengths, and state scope. The National Waterways Act, 2016 supplies the statutory boundary. Neither supplies current operational evidence or a reusable route dataset.

Spatial boundary: cached OpenStreetMap Nominatim results under ODbL supply only separately attributed representative navigation points on one named component. No point is an official Ministry coordinate, waterway centre, endpoint, surveyed site, channel limit, navigable extent, complete route, or proof of operation.

Entity boundary: National Waterway systems remain separate from development works on those systems. National Waterway 1 is linked to Jal Marg Vikas without merging identities or transferring lifecycle, money, schedule, organization, progress, or geometry claims.

D1 and serving boundary: the incremental idempotent migration adds 68 accepted records, the full 111-row ledger, two fingerprinted official sources, 204 source links, 136 typed observations, exact source-publisher roles, aliases, and the reviewed NW-1 relationship. Static browse and evidence reads remain the independent public fallback.

Rendered review: the exact production build was reviewed at 320, 390, 820 portrait, 1180 landscape, 1440, and 1920 pixels. The 5,287-record default non-toll directory, 72-record Logistics & Inland Transport facet, National Waterway search and Project Brief, clean Home reset, phone detail sheet, tablet selected workspace, and static map-failure access remained usable without page-level horizontal overflow or browser console errors.

Remaining gap: 5,287 core records represent 52.87% of the 10,000-record target. Forty-three National Waterway candidates still need stronger reusable geography, while permission-cleared railway, road, logistics, industrial, grid, digital, and environmental inventories remain the larger scale bottleneck.

Twelve operational major-port assets with dated cargo evidence

A sixth auditable source lane expands the public catalog to 5,308 source-backed records without redesigning the accepted map, discovery workspace, Project Brief, optional toll layer, or specialist evidence model.

Source boundary: the Ministry's Basic Port Statistics 2024-25 supplies the major-port identity and topography references. The May 2026 cargo report supplies advance-estimate tonnes and the dated operational observation. Neither source supplies a complete port boundary or development-project dossier.

Spatial boundary: the source's degree-minute-second coordinates were normalized into longitude-latitude order for WGS 84 display. The table does not publish its geodetic datum, and its points are not port limits, berth or terminal boundaries, cadastral parcels, surveyed centres, or complete asset extents.

Entity boundary: seven earlier port development projects remain separate from the 12 port assets. Five reviewed asset-to-development-project links are stored without merging terminal, berth, jetty, or pipeline works into their parent port asset.

Containment boundary: five official points fall inside the reviewed land polygons and one district. Seven coastal reference points sit outside the land-only state and district polygons, so the explicit coastal exception remains visible and the points were not moved inland.

D1 and serving boundary: the incremental idempotent migration adds the 12 accepted records, full 14-row candidate ledger, two immutable official documents, 24 source links, 24 field observations, exact authority roles, aliases, and reviewed relationships. Static browse and evidence reads remain the independent public fallback.

Rendered review: the exact production build was reviewed at 320, 390, 820 portrait, 1180 landscape, 1440, and 1920 pixels. The 5,219-record default non-toll directory, 19-record Port facet, Chennai Port search and Project Brief, clean Home reset, phone detail sheet, tablet selected workspace, and static map-failure access remained usable without page-level horizontal overflow or browser console errors.

Remaining gap: 5,219 core records represent 52.19% of the 10,000-record target. Large permission-cleared transport, logistics, industrial, grid, digital, and environmental inventories remain the scale bottleneck.

One hundred five licensed public-use airport assets

A fifth auditable source lane expands the public catalog to 5,296 source-backed records without redesigning the accepted map, discovery workspace, Project Brief, optional toll layer, or specialist evidence model.

Source boundary: the DGCA public-use register supplies aerodrome identity, ICAO indicator, licence number, validity dates, and the 17 July 2026 source update. It does not publish state, district, coordinate, geometry, operator, capacity, or present service.

Spatial boundary: 103 points come from a pinned public-domain OurAirports snapshot. Datia and Pantnagar use cached OpenStreetMap Nominatim results under ODbL. Every point remains representative, not a DGCA coordinate, aerodrome reference point, runway point, terminal site, parcel, or airport polygon.

Review boundary: Aizawl and Rewa remain held after their DGCA validity dates passed, Bhogapuram remains held without a licence number or validity dates, and Shravasti remains held because current DGCA and AAI licence identities conflict. Fifteen asset-to-project relationships remain separately reviewable and do not merge the entities.

D1 and serving boundary: an incremental idempotent migration adds the 105 accepted records and complete 110-row ledger. Twelve ordered query sets retain exact D1-to-static parity, while static browse and evidence reads remain the independent public fallback.

Rendered review: the exact production build was reviewed at 320, 390, 820 portrait, 1180 landscape, 1440, and 1920 pixels. The 5,207-record default non-toll directory, 130-record Airport facet, Juhu Airport search and Project Brief, clean Home reset, phone detail sheet, tablet selected-marker framing, and static map-failure access remained usable without page-level horizontal overflow or browser console errors.

Remaining gap: the catalog is still water-heavy. Permission-cleared, current transport, logistics, industrial, grid, digital, and environmental inventories with stable identities and reusable geography remain the primary bottleneck.

One hundred two historical generation assets from a fourth source lane

A conservative Central Electricity Authority batch expands the public catalog to 5,191 source-backed records without redesigning the map, discovery workspace, markers, Project Brief, optional toll layer, or specialist evidence model.

Source boundary: CEA supplies official station identity, state, fuel grouping, monitored capacity, and dated generation fields through data.gov.in under the Government Open Data License India. The source does not publish coordinates or current lifecycle.

Spatial boundary: Wikidata supplies only a frozen CC0 navigation point after conservative name matching, unique-item review, India bounds, and state containment. It is not an official CEA coordinate, surveyed facility point, plant boundary, unit point, or exact project footprint.

Deduplication boundary: hydro candidates remain deferred until they can be reconciled with existing NWIC powerhouse assets. Development projects and physical station assets remain different entity kinds; no relationship is merged merely from a similar name or nearby point.

D1 and serving boundary: an incremental idempotent migration adds the 102 accepted records and the complete 526-row control ledger. Twelve ordered query sets retain exact D1-to-static parity, while static browse and evidence reads remain the independent public fallback.

Remaining gap: the catalog is still water-heavy, and this CEA source is historical. Current, permission-cleared transport, logistics, industrial, grid, digital, and environmental inventories with stable identities and reusable geography remain the primary bottleneck.

Five thousand accepted core records with three auditable source lanes

Two thousand five hundred reviewed additions expand the public catalog to 5,089 source-backed records without changing the accepted map, Project Brief, optional toll layer, evidence semantics, or individual unclustered marker system.

Count boundary: the 5,089-record catalog contains exactly 5,000 accepted core projects or infrastructure assets and 89 separately reported toll-plaza facilities. The 2,500 additions are 1,873 NWIC, 309 PAIMANA, and 318 AMRUT records.

Review boundary: NWIC retains 3 duplicates, 184 needs-review rows, and 3,375 pending rows. PAIMANA retains 10 duplicates, 722 needs-review rows, and 94 rejected or deferred rows. AMRUT retains 18 anchor collisions for review and 8,398 pending rows.

Spatial boundary: all 5,089 positions pass independent state and district checks. Twenty-eight source-state differences remain explicit. Kakari Mahudi Dam is returned by both Arvalli and Mahisagar district boundaries, so no district is assigned to that record.

Serving boundary: the complete spatial index is about 8.4 MB uncompressed, while the first 24-row discovery response with 5,000 active IDs is about 130 KB. Removing a shared runtime import of the full source catalog reduces the compiled InfrastructureMap client chunk from 41.3 MB to about 227 KB. D1 imports the three source lanes idempotently and retains exact query parity; static read models remain the independent fallback.

Coverage limitation: all maintained categories and every state or union territory have records, but the checkpoint remains strongly water-heavy and many source dates are stale. Airports, ports, digital, logistics, industry, grid, pipelines, and other network categories need stronger current authoritative inventories before 10,000.

Two thousand five hundred core records and a second authoritative source family

One thousand five hundred eighty-nine reviewed National Water Informatics Centre assets expand the public catalog to 2,589 source-backed records without changing the accepted map, Project Brief, filters, optional toll layer, or deeper evidence model.

Source boundary: the three reviewed resources are the OGD Point Layer of Dams of India, Point Layer of Powerhouse, and Point of Barrage, Weir, or Anicut. Archive hashes, byte lengths, source fields, retrieval date, and the Government Open Data License India remain recorded.

Review boundary: coordinate disagreements, JK-coded geography requiring current J&K or Ladakh review, missing state scope, two border-sensitive positions, and unselected source rows remain outside the accepted count. Three source rows are linked as reviewed duplicates rather than double-counted.

Spatial boundary: all 2,589 points pass state and district containment, with 25 source-state differences kept explicit. Containment does not upgrade an official point into a surveyed footprint, reservoir polygon, route, or project extent.

Serving boundary: the 24-row first discovery response is 79.6 KB, under 2% of the full fallback fact index. When more than 1,200 records match, national zoom deterministically prioritizes 700 individual markers through a balanced established-project and baseline-asset mix, then reveals the remainder with zoom or filtering while the exact matching count stays visible. D1 stages both source control planes and twelve exact query checks; versioned static reads remain the independent public fallback.

Coverage limitation: the checkpoint is intentionally source-backed but water-heavy. The next 5,000-record phase must add more authoritative source families and improve category and state balance rather than treating this source distribution as national completeness.

One thousand reviewed records and an auditable D1 inventory control plane

Three hundred eighty-two reviewed PAIMANA projects expand the public catalog to 1,000 source-backed records. The accepted interface remains unchanged while the data path now separates broad baseline inventory from selective deep evidence.

Count boundary: 1,000 is the complete public catalog count. The 10,000-project objective is at 911 because the 89 reviewed toll plazas remain a separate facility layer.

Source boundary: all additions come from the official May 2026 PAIMANA candidate universe. Sixty-six replacement candidates passed independent state-matched geocoding after earlier candidates failed the spatial gate. No fallback coordinate was invented.

Serving boundary: the public map continues to use compact versioned spatial and discovery reads with static failure fallback. D1 now exposes the candidate ledger, maturity counts, and twelve exact ordered-query parity checks, but full public discovery cutover remains staged.

Next checkpoint: reach 2,500 accepted core projects using more than one major authoritative source family and materially broader category and state coverage.

One hundred more reviewed systems and twelve exact D1 query checks

One hundred reviewed central-sector projects expand the public catalog to 618 individual source-backed records. The batch preserves the exact first 370 PAIMANA selections and adds railway systems, national highways, public healthcare and higher-education infrastructure, fuel systems, and steel or mineral-processing development from the same May 2026 official source universe.

Selection boundary: these 100 records were selected from official sector and physical-scope evidence, reviewable named geography, state balance, and spatial-review readiness. They are not a cost ranking, a complete inventory, or a judgment about unselected projects.

Spatial boundary: representative facility, locality, and corridor anchors remain navigation context, not surveyed sites, alignments, rights of way, project extents, plant boundaries, mine leases, or operating evidence. Twenty-four new and 55 cumulative shared-anchor cases keep every reviewed coordinate unchanged.

Interface boundary: the accepted map, compact Project Brief, filters, optional toll layer, shareable state, and specialist evidence tabs are unchanged. The larger catalog uses the same points-first read path and native individual MapLibre features.

Rendered release gate: the exact build passed the 320, 390, 820, 1180, 1440, and 1920 pixel review with no page overflow, working selected-record and failure paths, and a measured 142 ms first-marker point on local desktop. The five-million-row local SQLite metadata benchmark retained single-digit-millisecond query times; neither measurement is presented as production D1 latency.

Still incomplete: 1,517 PAIMANA source rows remain unselected. Many need stronger entity resolution, supported geometry, or package-versus-project review. D1 remains a shadow read model; R2, PMTiles, full-text search, production-like concurrency evidence, and a public database read cutover remain future scale work.

Fifty more reviewed systems and exact D1 query parity

Fifty reviewed central-sector projects expand the public catalog to 518 individual source-backed records. The batch preserves the exact first 320 PAIMANA selections and strengthens mine development, ports and fuel systems, environmental and water utilities, public campuses, generation, transmission, railways, and highways from the same May 2026 official source universe.

Selection boundary: these 50 records were selected for physical-system scope, reviewable named geography, and spatial-review readiness rather than cost. Previously curated projects, note-named source conflicts, ambiguous multi-state systems, and records without a truthful supported map representation remain outside the batch.

Spatial boundary: representative localities and corridor anchors remain navigation aids, not surveyed sites, alignments, midpoints, rights of way, or project extents. Eight new shared-anchor cases and 31 cumulative collisions retain every original coordinate and an explicit review; no jitter or invented displacement is used.

Read-path boundary: D1 remains a shadow read model. The public map, directory, Project Brief, and evidence views continue to use validated versioned artifacts and remain available if D1 fails. A public D1 cutover still requires database-backed cursor parity, rollback testing, failure isolation, and production latency evidence.

Still incomplete: the 1,987-row PAIMANA source universe still contains 1,617 unselected rows. Many are package-level highway or rail records, ambiguous multi-state systems, or records needing stronger identity and geography review. R2, PMTiles, full-text search, and production-like D1, tile, and concurrency load testing remain future scale work.

Twenty-five more reviewed systems and semantic D1 parity

Twenty-five reviewed central-sector projects expand the public catalog to 468 individual source-backed records. The batch preserves the exact first 295 PAIMANA selections and adds aviation, generation, grid, metro, environmental utility, water-supply, public-campus, medical-college, and fuel-network records from the same May 2026 official source universe.

Selection boundary: these 25 records were selected for physical-system scope, reviewable named geography, and spatial-review readiness rather than cost. Known duplicate base projects and broad multi-state rows without a truthful map representation remain outside this batch.

Spatial boundary: representative localities and corridor anchors remain navigation aids, not surveyed sites, alignments, midpoints, rights of way, or project extents. Two new shared-anchor cases and 23 cumulative collisions retain every original coordinate and an explicit review; no jitter or invented displacement is used.

Rendered review: the exact production build was reviewed at 320, 390, 820 portrait, 1180 landscape, 1440, and 1920 pixels. The 379-record default directory, individual map features, new AIIMS Bibinagar Project Brief, phone detail sheet, selected-marker framing, and explicit basemap-failure directory remained usable without page-level horizontal overflow or browser console errors.

Scale boundary: D1 remains a shadow read model rather than the source of public facts. Field-level query parity, database-backed cursors, rollback testing, production latency evidence, R2, PMTiles, full-text search, and viewport activation remain future work.

Fifty more major systems and a staged D1 parity check

Fifty reviewed central-sector projects expand the public catalog to 443 individual source-backed records. The batch preserves the exact first 245 PAIMANA selections and adds airports, generation and grid systems, metros, urban water and environmental utilities, public campuses and hospitals, pipelines, rail corridors, highways and bridges, and water-resource systems from the same May 2026 official source universe.

Selection boundary: these 50 records were selected for physical-system scope, sector balance, reviewable named geography, and spatial-review readiness rather than cost. Omission does not imply lower importance.

Spatial boundary: representative localities and corridor anchors remain navigation aids, not surveyed sites, alignments, midpoints, rights of way, or project extents. Eight shared anchors retain every original coordinate and an explicit collision review; no jitter or invented displacement is used.

Scale boundary: the public discovery and detail APIs still read validated release artifacts. D1 is a staged semantic-parity path, not yet the source of public project facts; R2, PMTiles, database cursors, production query percentiles, and viewport activation remain future work.

Polavaram audit, contract, and funding trail

The 393-record catalog remains unchanged while the Polavaram Irrigation Project gains six official CAG and Ministry of Jal Shakti sources, six structured audit observations, four separately typed money records, three historical construction-contract records, seven exact organization-role rows, and three additional milestones. The record moves from Timeline validated to Contract-linked without changing its representative Polavaram locality point or current PAIMANA facts.

Audit-time boundary: the CAG audit is not a current status report. Its fieldwork ran from June 2022 to June 2023, its performance scope ends in March 2023, and government replies extend through November 2024.

Finding boundary: the interface says what the CAG reported, calculated, recommended, and answered after the government reply. It does not convert an audit observation into a product-authored allegation of corruption, misconduct, blame, or present failure.

Source-access boundary: the official CAG landing page and documents were reviewed through the public document viewer, but the CAG host did not permit a stable local byte retrieval during this release. The PDFs therefore have page-level review notes but no claimed local SHA-256 fingerprint.

Still incomplete: a complete current contract register, signed contracts and amendments, certified bills and payments, rehabilitation outcomes, current component status, final account, actual or final cost, and authoritative project geometry remain unlinked.

Bikaner IV Part A procurement, licence, and monitoring trail

The 393-record catalog remains unchanged while the Bikaner IV renewable-energy transmission Part A scheme gains six fingerprinted official sources, four procurement, agreement, tariff, or licence records, thirteen dated milestones, seven exact organization-role rows, and six CEA-monitored components. The record moves from Timeline validated to Contract-linked without changing its representative Bikaner coordinate.

Licence boundary: the final inter-State licence does not prove EPC award, construction completion, energisation, commissioning, present operation, tariff recovery, or final cost.

Schedule and entity boundary: Amendment VI's planned dates and CERC's recorded events remain separate. Part A remains distinct from the separately published Siwani Part B project, SPV, cost, tariff, licence, and monitoring row.

Spatial boundary: the unchanged Bikaner marker remains a representative corridor anchor, not the pooling station, a transmission-line route, right of way, surveyed site, or project extent.

Still incomplete: signed agreements, SPV acquisition price, EPC procurement and award, payments, invoices, amendments, disputes, audits, actual completion, operation, actual cost, final cost, authoritative geometry, and status after May 2026 remain unlinked.

Bikaner REZ Phase IV Part B procurement, licence, and monitoring trail

The 393-record catalog remains unchanged while the Siwani transmission scheme gains five fingerprinted official sources, four procurement, agreement, tariff, or licence records, twelve dated milestones, seven exact organization-role rows, and five CEA-monitored components. The record moves from Timeline validated to Contract-linked without changing its representative Siwani coordinate.

Licence boundary: the final inter-State licence does not prove EPC award, construction completion, energisation, commissioning, present operation, tariff recovery, or final cost.

Schedule and spatial boundary: November 2026 and June 2027 remain CEA target months, not an inferred breach or blame finding. The unchanged Siwani marker remains a representative corridor anchor, not a substation site, line route, right of way, or project extent.

Still incomplete: signed agreements, SPV acquisition price, EPC procurement and award, payments, invoices, amendments, disputes, audits, actual completion, operation, actual cost, final cost, authoritative geometry, and status after May 2026 remain unlinked.

Dhubri-Phulbari bridge approval, procurement, award, and financing trail

The 393-record catalog remains unchanged while the Dhubri-Phulbari Brahmaputra Bridge gains seven reviewed official sources, one construction-award record, nine dated milestones, and four exact organization-role rows. The record moves from Timeline validated to Contract-linked, adds Meghalaya to its official discovery scope, and keeps the existing Dhubri representative corridor point unchanged.

Progress and freshness boundary: PAIMANA reports 76.01% physical progress by 24 June 2026. NHIDCL's current page separately publishes 76% physical and 74% financial progress but no row-level status date, so those values remain retrieval-dated.

Procurement boundary: the first bidding process was annulled after one responsive bid; the re-invited cycle later reached technical evaluation and award. These facts do not establish fault, corruption, payment, final cost, certified completion, opening, or contractor performance.

Still incomplete: the signed construction contract, amendments, payment certificates, final account, certified completion, opening to traffic, current row date, official alignment geometry, disputes, audits, and authoritative corruption findings remain unlinked.

Dholera airport approval, package-award, and status trail

The 393-record catalog remains unchanged while Dholera International Greenfield Airport gains six reviewed official sources, three package contract awards, eight dated milestones, and eight exact organization-role rows. The record moves from Timeline validated to Contract-linked without changing its OpenStreetMap facility representative point.

Money boundary: the ₹636.4454249 crore, ₹333 crore, and ₹18.9880385 crore package awards are not summed as airport cost. PAIMANA's ₹922.78 crore Cumulative Expenditure remains separate from approved cost, payment, actual cost, and final cost.

Schedule and responsibility boundary: past package schedule dates do not prove breach, delay cause, or contractor fault. DIACL is the Cabinet-named implementing company; the three contractors retain package-specific roles.

Still incomplete: signed contracts, amendments, payment records, certified completion, DGCA licensing, opening, present operation, final accounts, disputes, audits, final cost, and official package geometry remain unlinked.

Bidar two-cycle procurement, amendment, and tariff trail

The 393-record catalog remains unchanged while the Bidar 2,500 MW solar-zone transmission record gains eight fingerprinted official sources, four procurement or agreement records, fourteen dated milestones, nine exact organization-role rows, and two CEA-monitored components. The record moves from Timeline validated to Contract-linked without changing its representative Bidar coordinate.

Agreement boundary: CERC records the ₹85.90 crore Contract Performance Guarantee, Share Purchase Agreement, SPV acquisition, and tariff adoption subject to licensing. The order is not the signed agreement text, acquisition price, project payment, or proof of present operation.

Schedule and spatial boundary: February 2026 and August 2026 remain source-published target months, not an inferred contractual breach or blame finding. The unchanged Bidar point remains a representative corridor anchor, not the pooling station, line alignment, midpoint, right of way, or project extent.

Still incomplete: signed agreement text, acquisition price, final licence order, complete EPC packages, payments, tariff recovery, disputes, audits, actual cost, final cost, authoritative geometry, and status after May 2026 remain unlinked.

Bhadla-Bikaner procurement, tariff, and component trail

The 393-record catalog remains unchanged while the Bhadla-III and Bikaner-III interconnection-strengthening record gains four fingerprinted official sources, three procurement or agreement records, thirteen dated milestones, six exact organization-role rows, and one CEA-monitored line component. The record moves from Timeline validated to Contract-linked without changing its representative Bhadla coordinate.

Money boundary: the ₹901.41 crore CERC project estimate, ₹1,382 crore CEA and PAIMANA scheme cost, ₹97.97 or ₹97.99 crore-per-year charges, and ₹30 crore guarantee remain differently defined quantities. No compatible project-cost revision is inferred.

Schedule and spatial boundary: August 2026 and March 2027 remain published schedule labels, not an inferred contract amendment or blame finding. The unchanged Bhadla point remains a representative corridor anchor, not a pooling station, line alignment, midpoint, right of way, or project extent.

Still incomplete: complete RFP and signed agreement text, acquisition price, amendments, final transmission-licence order, EPC packages, payments, tariff recovery, disputes, audits, actual cost, final cost, authoritative geometry, and status after May 2026 remain unlinked.

South Olpad procurement, tariff, and component trail

The 393-record catalog remains unchanged while the Khavda Phase-IV 7 GW Part B transmission record gains three fingerprinted official sources, three procurement or agreement records, nine dated milestones, six exact organization-role rows, and four CEA-monitored components. South Olpad moves from Timeline validated to Contract-linked without changing its representative Olpad coordinate.

Source difference: CERC paragraph 29 reproduces text calling the selected tariff lower than a numerically smaller levelized tariff, while adjacent paragraphs describe it as 16.33% higher. The inconsistency remains explicit and does not become a project-cost conclusion.

Schedule and spatial boundary: CEA's October 2027 anticipated target is not relabeled as an executed contractual extension or delay finding. The unchanged Olpad point remains a representative corridor anchor, not South Olpad GIS, a line alignment, right of way, or project extent.

Still incomplete: signed agreement text, complete amendments, final transmission-licence order, EPC packages, payments, tariff recovery, disputes, audits, actual cost, final cost, authoritative geometry, and status after May 2026 remain unlinked.

Power-grid breadth and Koppal–Gadag component monitoring

Twenty additional Power Grid & Storage systems expand the source-backed national catalog to 393 individual records while preserving the 89-plaza opt-in layer and every earlier evidence trail. Koppal–Gadag also gains a fingerprinted CEA monitoring source with seven named scheme components and source-literal construction indicators.

Compatible scheme fields: CEA publishes ₹4,445 crore as Estimated cost, December 2023 as SPV transfer, December 2025 as Original SCOD, and September 2026 as Anticipated SCOD. These align with the compatible PAIMANA scheme snapshot but are not actual or final cost, payment, expenditure, or proof of completion.

Component boundary: “Stringing”, foundation, tower-erection, land-acquisition, or targeted-charge fields describe the named component in the reviewed CEA row. Release 54 does not turn them into a percent, cause, verdict, contractor performance claim, or whole-scheme state.

Still incomplete: surveyed line geometry, substation coordinates, package tenders, signed contracts, payments, amendments, disputes, audits, actual outturn, final cost, and newer component status remain unlinked.

KLI agreement trail and ten-second Project Brief

The Kochi–Lakshadweep Islands submarine OFC record now links two official, page-reviewed agreements and moves from Timeline validated to Contract-linked. The selected-project Overview is also rebuilt as a shorter, type-aware first read while the public catalog remains 373 records, the 284-record default non-toll scope and 89 optional toll plazas remain unchanged, and every source coordinate and precision label is preserved.

Money boundary: the 2022 agreement's ₹810.27 crore funding estimate, DBN's ₹837 crore capital-expenditure field and ₹1,072 crore Total financial implication, the 2024 ₹51.05 crore annual estimate, extra PMC, actual-based claims and DBN's ₹235 crore five-year OPEX use different definitions. They are not silently combined into a revision or payment series.

Spatial and schedule boundary: agreement clauses do not publish a surveyed cable route, landing-station coordinate set, authoritative actual completion date, or revised deadline. Kavaratti remains a representative corridor navigation point, and no delay is calculated by comparing a rollout clause with the later inauguration event.

D1 staging: the idempotent import retains two KLI contract entities, one O&M relationship, two fingerprinted source documents and ten exact project-role rows. The public Site continues to use the validated static read path; D1 and R2 remain staged capability foundations.

Still incomplete: complete claims, advances, disbursements, vendor payments, annual O&M outturns, audit results, reconciliations, recoveries, final adjustments, present network performance, actual cost, final cost, disputes, audit findings and authoritative cable geometry remain unlinked.

CANI implementation, amendment, and O&M trail

The Chennai–Andaman and Nicobar Islands submarine OFC record now connects its parent-system evidence to four official, page-reviewed agreement documents. CANI moves from Timeline validated to Contract-linked while the public catalog remains 373 records, the 284-record default non-toll scope remains unchanged, and its representative Port Blair navigation anchor is preserved.

Source difference: the 2018 agreement's ₹929.358 crore USOF CAPEX ceiling, the later DBN page's ₹947.283 crore capital expenditure, and its ₹1,224 crore Total project cost are dated, differently labeled quantities. Release 52 does not silently convert them into one original-versus-revised cost series.

Read model: the staged D1 import now retains two agreement entities, two amendment entities, their two explicit relationships, four fingerprinted source documents, and seven exact CANI organization-role rows. The public Site continues to use the validated static read path; D1 and R2 remain staged capability foundations.

Home and return state: the complete brand block remains the explicit clean-root Home link. Pointer activation or unmodified Enter clears filters, questions, toll opt-in, selection, dossier tab, workspace, result position, and camera; refresh, direct links, shared URLs, modified new-tab activation, and Back or Forward continue to preserve their encoded browser state.

Still incomplete: a complete claims, disbursement, vendor-payment, annual O&M outturn, reconciliation, recovery, or final-adjustment ledger is not linked. Actual or final project cost, current operating performance, payment audit, dispute records, audit findings, and authoritative cable geometry remain missing.

Digital Bharat Nidhi submarine-fibre systems

Digital & Telecom is now an active evidence-backed category. Two major physical-network records add the Chennai–Andaman and Nicobar Islands submarine OFC system and the Kochi–Lakshadweep Islands submarine OFC system, bringing the public catalog to 373 individual records: 249 Major projects or mine assets and 124 optional Facilities or local schemes.

Default discovery: a fresh browse view now shows 284 non-toll records. The same 89 toll plazas remain in the 373-record catalog and appear only through the explicit Toll plazas layer.

Spatial review: independent state and district containment covers all 373 points. Three hundred sixty-seven match the declared state scope, the same six NHAI conflicts remain explicit, and every point resolves to one of 251 district-code pairs without exporting boundary geometry.

Still incomplete: no reviewed authoritative route LineString or landing-station coordinate register is published. Clause-level agreements, amendments, payments, disputes, actual completion costs, final costs, and current network availability remain future depth work.

Tech Atlas basemap parity, 16 infrastructure systems, and sourced inauguration

Infra now uses the complete OpenFreeMap Positron style instead of deleting most provider labels and restyling the map through a warm India wash. Roads, railways, airports, waterways, land cover, buildings, cities, towns, villages, and local labels disclose progressively with zoom while the reviewed India outline, state boundaries, all 371 individual records, and project evidence remain separate and preserved. Only country-label layers are narrowed to India so large neighboring-country names do not compete with the map's subject.

Evidence boundary: inauguration is stored with its date, scope, source, reporting cutoff, and named inaugurating officeholder only where the official report publishes them. It does not imply full-project construction completion, commissioning, opening, 100% progress, or current operation.

Default discovery: a fresh browse view now shows 282 non-toll records. All 89 reviewed toll-plaza records remain in the unchanged 371-record catalog and appear only when the accessible Toll plazas layer is enabled. The toggle composes with search, questions, categories, lifecycle, geography, evidence filters, results, markers, and shared URLs; a named toll search offers a direct enable-layer action, and direct toll-project URLs still open.

Home behavior: the complete India Infra Map brand lockup is a semantic Home link. Normal activation performs a clean-root navigation that clears discovery, dossier, workspace, result-position, and camera state; modified clicks retain standard open-in-new-tab behavior, while refresh, Back and Forward, direct project links, and shared stateful URLs keep their encoded restoration behavior.

Rendered reference review: an anonymous public browser check on 16 July 2026 reached Tech Atlas with the title “Tech Atlas | 581 reviewed Bay Area tech companies” and no login gate. Initial, wheel, pan, zoom, reset, hover, focus, selection, right-workspace, phone, tablet, laptop, wide-screen, and blocked-basemap states were compared with Infra. The provider palette, disclosure hierarchy, full-bleed composition, and restrained 410-pixel desktop workspace align; Infra deliberately keeps smaller precision-aware native features, India overlays, URL state, stronger keyboard preview, and an explicit map-failure directory instead of copying Tech Atlas's 581 DOM logo markers or Bay Area semantics.

Coverage boundary: Release 50 preserves all 371 records and adds no weak breadth record merely to populate a new category. Digital & Telecom has an icon and schema identity but is not offered as an active filter until reviewed records exist. Space and sensitive strategic infrastructure remain deferred.

Two hundred twenty-five reviewed PAIMANA projects

Twenty-five additional source-backed Major projects bring the public catalog to 371 individual records: 247 Major projects or mine assets and 124 optional Facilities or local schemes. The exact first 200 PAIMANA selections remain unchanged, and the ninth batch broadens public-system, institution, industrial, energy-network, rail, metro, highway, hydropower, and mining-development coverage.

Source freshness: the MoSPI release calendar schedules the June 2026 Flash Report for 25 July 2026, so the May 2026 report remains the newest published national PAIMANA breadth source reviewed for this 16 July release.

Still incomplete: these breadth records do not yet provide complete tender, award, signed contract, contractor, payment, amendment, dispute, audit, final-cost, or official project-geometry trails. The 225-record lane is a reviewed sample, not full PAIMANA or national coverage.

Smaller project medallions and progressive India context

The national map now gives every published record a compact circular category medallion while restoring useful Positron road, transport, water, land-cover, built-environment, and place context progressively by zoom. All 346 records, source coordinates, precision meanings, filters, lenses, URLs, Project Briefs, and evidence trails remain unchanged.

Orientation boundary: OpenStreetMap and OpenMapTiles roads, airports, stations, waterways, buildings, and place labels help people orient themselves. They do not establish a monitored project, project geometry, construction state, cost, deadline, contractor, progress, or evidence tier.

Collision boundary: every active record remains an individual coordinate-locked feature without clustering, count pills, jitter, or invented displacement. Zoom remains the current separation path for dense locations; deterministic cycling or click expansion is still future work.

Chennai Metro Phase II contract trail

Chennai Metro Phase II now connects its April 2026 PAIMANA progress, cost, and schedule snapshot to CMRL's current project status, archived and current awarded-tender registers, and two signed-contract releases. The record moves from Basic official record to Contract-linked while preserving all 346 published records and the existing Chennai locality representative point.

Source difference: CMRL's archive prints ₹91,85,06,227.23 and ₹89,85,06,227.23 for the same ASA07 tender reference, a ₹2 crore difference. No reviewed source labels either value a revision or superseding amount, so both remain visible and unresolved.

Still incomplete: the complete 60-row package register, consistent package values, payments, amendments, variations, claims, disputes, audits, final package values, current or final project cost, and official corridor geometry remain unlinked. A signed agreement does not prove present-day performance or completion.

Mumbai–Ahmedabad High-Speed Rail contract trail

The 508 km Mumbai–Ahmedabad High Speed Rail record now connects its April 2026 PAIMANA progress, cost, and schedule snapshot to NHSRCL's official package structure, aggregate civil-award release, agreement records, and a bounded selection of core awarded-tender rows. The project moves from Basic official record to Contract-linked while keeping all 346 records, the existing representative corridor point, and every earlier evidence trail unchanged.

Source evolution: NHSRCL's package-structure page is marked 13 August 2019. Later dated award and agreement sources supersede that snapshot's procurement state only for the facts they publish. Bidder and technical-bid pages are not treated as award proof.

Still incomplete: package-specific award dates, agreements, organizations, and values remain incomplete across the reviewed slice. A complete payment, amendment, variation, dispute, audit, and final-cost trail is not linked, and the map position remains a representative corridor point rather than official alignment geometry, package extent, station, or work site.

Mumbai Metro Line 3 contract trail

Mumbai Metro Line 3 now connects its PAIMANA progress, cost, expenditure, and schedule snapshot to MMRCL's official package-status, funding, consultant, and operations work-order pages. The record moves from Timeline validated to Contract-linked while keeping its representative corridor point and every earlier catalog record unchanged.

Funding boundary: MMRCL publishes ₹23,136 crore as Total Project Completion Cost, while its eight displayed funding components sum to ₹23,137 crore and 100%. The ₹1 crore displayed difference remains visible. This funding plan is not substituted for PAIMANA's ₹37,276 crore Revised Cost or ₹33,648.26 crore Cumulative Expenditure.

Still incomplete: most package award dates, nine package amounts, package payment series, signed contract files, amendments, variations, disputes, audits, current operations performance, and final package or project costs remain unlinked. The work-order register does not prove present-day service status or full-project completion.

Six evidence questions with exact reasons

The browse workspace can now answer six deterministic, shareable public-awareness questions: airports in progress, explicit high-speed rail, past reported deadlines, completion dates moved by at least ten years, compatible revised costs at least twice the original, and records carrying an explicit freshness warning. Every result states why it matched, while the question summary exposes the scope, required fields, denominator, missing-data count, and catalog date.

Current answers: 17 airports report progress between 1% and 99%; one reviewed record is explicitly high-speed rail; 22 projects had passed their reported deadline in the same status snapshot; five completion dates moved by at least ten years; 16 compatible revised costs are at least twice their original; and 29 records carry an explicit freshness warning.

Measured local check: an unthrottled cache-disabled 390 by 844 run showed the source-backed overview at 24.5 ms, browse controls at 63.0 ms, all 346 native project features at 160.7 ms, orientation geography at 226.9 ms, FCP and LCP at 56 ms, zero layout shift, no long task, and about 33.3 MB used JS heap. The comparable Release 43 run recorded 18.6, 74.8, 119.8, and 194.3 ms, FCP and LCP at 72 ms, and about 61.3 MB. A 1.4-second map-drag sample reported 60.1 fps, 17.7 ms p95 frame time, and no frame over 50 ms. Twenty sequential warm local question requests measured 1.3 ms p95. These are controlled local observations, not production percentiles; the later marker and geography times remain within the interaction session but are modestly slower than the earlier single run.

Interpretation boundary: these are source-bounded comparisons, not findings about cause, fault, delay responsibility, corruption, waste, success, current operation, or present-day status. A record without a freshness warning is not certified current or complete.

Two hundred reviewed PAIMANA projects

Twenty-five more source-backed Major projects bring the public catalog to 346 individual records: 222 Major projects or mine assets and 124 optional Facilities or local schemes. The first 175 PAIMANA records remain in their exact published order. The new batch adds airports, metros, rail, a Brahmaputra bridge, highways, hydro and power systems, energy pipelines, major water systems, a national public-university campus phase, and coal development without changing the accepted map, evidence, search, Project Brief, URL, accessibility, or fallback behavior.

Measured local check: under the same cache-disabled 1.6 Mbps, 150 ms latency, 4-times CPU, 390 by 844 profile, the 346-record overview appeared at 3.40 seconds, browse controls at 9.51 seconds, all native project features at 12.16 seconds, and orientation geography at 15.34 seconds, with zero layout shift and no long task. The Release 42 321-record values were 3.40, 9.12, 11.95, and 15.14 seconds. An unthrottled cache-disabled run showed the overview at 18.6 ms, browse controls at 74.8 ms, all 346 project features at 119.8 ms, FCP and LCP at 72 ms, zero layout shift, no long task, and about 61.3 MB used JS heap. A 1.2-second map-drag sample reported 61.2 fps, 17.5 ms p95 frame time, and no frame over 50 ms. The constrained Tirap Project Brief became visible in 3.06 seconds; its brief API request took 178 ms. These are controlled local observations, not production percentiles.

Still incomplete: these PAIMANA breadth records do not yet provide complete tender, award, signed contract, contractor, payment, audit, dispute, final-cost, or project-geometry trails. The 200-record lane is a reviewed sample, not full PAIMANA or national coverage.

One hundred seventy-five reviewed PAIMANA projects

Twenty-five more source-backed Major projects bring the public catalog to 321 individual records: 197 Major projects or mine assets and 124 optional Facilities or local schemes. The first 150 PAIMANA records remain in their exact published order. The new batch adds airports, metros, rail, Ganga bridges, highways, hydro and power systems, coal and mining development, and major irrigation systems without changing the accepted map, evidence, search, Project Brief, URL, accessibility, or fallback behavior.

Measured local check: under the same cache-disabled 1.6 Mbps, 150 ms latency, 4-times CPU, 390 by 844 profile, the 321-record overview appeared at 3.40 seconds, browse controls at 9.12 seconds, all native project features at 11.95 seconds, and orientation geography at 15.14 seconds, with no long task. The Release 41 296-record results were 3.73, 8.66, 11.70, and 14.88 seconds respectively. An unthrottled local run showed the overview at 46.5 ms, all 321 project features at 209 ms, FCP and LCP at 164 ms, CLS 0.1, and one 74 ms startup long task; an 820-pixel selected-view drag sample reported 60 fps, 18 ms p95 frame time, and no frame over 50 ms. These are controlled local observations, not production percentiles.

Still incomplete: these PAIMANA breadth records do not yet provide complete tender, award, signed contract, contractor, payment, audit, dispute, final-cost, or project-geometry trails. The 175-record lane is a reviewed sample, not full PAIMANA or national coverage.

Real project geography appears before the interactive map is ready

The catalog, source coordinates, spatial precision, evidence, Project Briefs, filters, URLs, and individual MapLibre features remain unchanged. Startup now leads with a build-generated national overview containing all 296 published records at their reviewed coordinates while the interactive renderer and directory finish loading.

Measured local change: under the same cache-disabled 1.6 Mbps, 150 ms latency, 4-times CPU, 390 by 844 profile, the truthful overview appeared at 3.73 seconds and all 296 native project features at 11.70 seconds, down from 22.44 seconds. Browse controls improved from 9.40 to 8.66 seconds, orientation geography from 25.45 to 14.88 seconds, and the run recorded no long task. An unthrottled 1,440 by 900 drag and wheel sample reported 60.7 fps with no frame over 50 ms. These are controlled local observations, not production percentiles.

Still incomplete: the constrained native-feature time remains above the aspirational 3-second mobile budget, and the pinned renderer is still the largest startup asset. Production p75 evidence, D1 reads, R2 or PMTiles delivery, viewport activation for much larger catalogs, and the next evidence expansion remain staged.

One hundred fifty reviewed PAIMANA projects

Twenty-five more source-backed Major projects bring the public catalog to 296 individual records: 172 Major projects or mine assets and 124 optional Facilities or local schemes. The first 125 PAIMANA records remain in their exact published order. The sixth batch adds coverage across 11 official sectors and 14 state or union-territory scopes, with particular attention to Northeast, Himalayan, Punjab, energy, minerals, water, bridge, road, rail, and public-institution systems.

Performance check: a cached local production run showed all 296 markers in 386.1 ms and a 61.2 fps drag sample with no frame over 50 ms. A cache-disabled 1.6 Mbps and 4-times CPU mobile profile remained over budget, with the spatial index at 9.54 seconds and first markers at 22.46 seconds. These are local observations, not production percentiles; the constrained startup path remains a stated scale limit.

Still incomplete: these PAIMANA breadth records do not yet provide complete tender, award, signed contract, contractor, payment, audit, dispute, final-cost, or project-geometry trails. The 150-record lane is a reviewed sample, not full PAIMANA or national coverage.

One hundred twenty-five reviewed PAIMANA projects

Twenty-five more source-backed Major projects bring the public catalog to 271 individual records: 147 Major projects or mine assets and 124 optional Facilities or local schemes. The first 100 PAIMANA records remain in their exact published order, while the accepted map, evidence, search, Project Brief, specialist, URL, accessibility, and fallback behavior remains intact.

Still incomplete: these PAIMANA breadth records do not yet provide complete tender, award, signed contract, contractor, payment, audit, dispute, final-cost, or project-geometry trails. The 125-record lane is a reviewed sample, not full PAIMANA or national coverage.

The PAIMANA Major lane reaches one hundred reviewed projects

Twenty-five more source-backed Major projects bring the public catalog to 246 individual records: 122 Major projects or mine assets and 124 optional Facilities or local schemes. Content-addressed read models and release-versioned API requests ensure returning browsers retrieve one matching catalog instead of combining old and new cached artifacts. The accepted map, evidence, search, Project Brief, specialist, URL, accessibility, and fallback behavior remains intact.

Still incomplete: these PAIMANA breadth records do not yet provide complete tender, award, signed contract, contractor, payment, audit, dispute, final-cost, or official project-geometry trails. The hundred-record lane is a reviewed sample, not full PAIMANA or national coverage.

Seventy-five PAIMANA projects widen the Major layer

Twenty-five source-backed Major projects extend the reviewed PAIMANA lane without changing its first 50 records. The public catalog now contains 221 individual records: 97 Major projects or mine assets and 124 optional Facilities or local schemes.

Still incomplete: PAIMANA breadth records do not yet provide complete tender, award, signed contract, contractor, payment, audit, dispute, or official project-geometry trails. These gaps remain visible, and the existing deep-dossier lane continues separately.

The directory now loads one evidence-aware result page at a time

All 196 reviewed records, individual map features, source coordinates, precision, evidence, Project Briefs, filters, URLs, and public access remain unchanged. The healthy browse path no longer downloads the complete 316,598-byte current-facts artifact before the directory becomes useful.

Still incomplete: the API reads the versioned release artifact rather than D1, and returning all matching IDs is appropriate for 196 records but not a million-record viewport. Database cursors, full-text search, viewport activation, PMTiles, immutable hosting caches, production concurrency evidence, and new project coverage remain staged.

Official map context now arrives in a compact verified package

All 196 reviewed records, source coordinates, precision, evidence, Project Briefs, interaction, and public access remain unchanged. The reviewed India outline, state boundaries, and state labels now hydrate from three content-addressed gzip assets with independent raw-source fallbacks.

Still incomplete: this is a compact release artifact, not PMTiles or a production D1 or R2 read path. The 196-record current-fact index still loads in the browser, Evidence remains the largest specialist response, and this release adds no new project coverage.

Specialist evidence now loads one section at a time

All 196 reviewed records, evidence semantics, coordinates, map interaction, Project Briefs, and public access remain unchanged. Timeline, Money, Organizations, Packages, Contracts, and Evidence now have independently cached specialist sections.

Still incomplete: Evidence remains the largest specialist response because nested source fields are restored across the dossier. The current-fact index and raw geography remain larger browser payloads, D1 and R2 are still unbound, and this release adds no new project coverage.

The Project Brief now arrives before specialist evidence

All 196 reviewed records, evidence ledgers, coordinates, precision, map layers, discovery state, and public access remain unchanged. Selecting a record now retrieves a compact Project Brief first. Timeline, Money, Organizations, Packages, Contracts, and Evidence are retrieved only after a reader opens a specialist tab.

Still incomplete: the first specialist action currently retrieves one cached investigation bundle, so later specialist tabs do not make another request but are not yet independently partitioned. The current-fact index and raw geography remain larger browser payloads, D1 and R2 are still unbound, and this release adds no new project coverage.

The map now reads like an infrastructure atlas before a reader opens a record

The 196 reviewed records, evidence ledgers, coordinates, precision, native MapLibre delivery, selected-record API, search, filters, URLs, Project Briefs, and public access are unchanged. The visual system now follows the approved pale-water, warm-land, compact-command, multicolor-medallion, and answer-first dossier direction without importing any illustrative value from the art reference.

Still incomplete: the visual hierarchy does not solve current geographic density at a few national-scale hotspots, create missing line or area geometry, activate D1 or R2, split specialist-tab payloads, or add a submission workflow. All 196 active records remain individual by design, so overlap is resolved through zoom and filtering rather than count clusters or coordinate movement.

Real project points arrive first, and the catalog no longer ships inside the app bundle

The 196 Release 30 records, evidence ledgers, coordinates, precision, categories, filters, Project Briefs, and public access remain unchanged. A compact spatial index now hydrates before current discovery facts, while selecting a record retrieves that one dossier through a stable versioned API path.

Still incomplete: local measurements are not p75 field data or production D1 latency. Current geography remains a large raw payload, initialization still records long tasks above 50 ms, specialist tabs still co-travel with the selected dossier, D1 and R2 are not bound, and vector tiles, cursor pagination, list virtualization, immutable hosting caches, slow-network evidence, and production-like concurrency remain staged work.

Seven mine assets across five states make the IBM breadth lane genuinely comparative

The default map now contains 196 source-backed records: 72 Major projects or mine assets and 124 optional Facilities or local schemes. Six additional Indian Bureau of Mines inspection records add bauxite, chromite, manganese, gold, copper, and lead-zinc assets across Chhattisgarh, Jharkhand, Odisha, Rajasthan, and Telangana while preserving the Release 29 Vijay-II iron-ore record.

Still incomplete: seven records are not a national mine inventory. Current operating state, lease polygons, extraction footprints, post-inspection production, reserves, revenue, tenders, contracts, payments, clearances, closure plans, decommissioning, and audit findings remain unlinked unless a reviewed source supports them.

Project types become explorable, and mines become their own evidence entity

The default map now contains 190 source-backed records: 66 Major projects or mine assets and 124 optional Facilities or local schemes. A shareable multi-select taxonomy lets readers isolate 14 reviewed categories, including dams, airports, highways, rail, metro, bridges, ports, energy, water, industrial development, and Mining and Minerals without classifying records from cost or project name alone.

Still incomplete: this is a one-record mine-asset lane, not a national mine inventory. Current mine operating status, lease polygons, post-2024 production, reserves, revenue, clearances, closure status, and audits remain unlinked. Municipal boundaries are not republished because the reviewed official district service prohibits offline export; the map keeps progressive capital, city, town, and village orientation without inventing civic geometry.

Major-project coverage reaches 33 state and union-territory scopes

A second reviewed PAIMANA breadth batch adds 25 source-backed Major projects without changing or replacing the first batch. The cumulative lane now contains 50 PAIMANA projects across 33 state or union-territory scopes and 11 official sectors, taking the public directory to 189 records: 65 Major projects and 124 optional Facilities or local schemes.

Boundaries of this release

• PAIMANA Original Cost, Anticipated Cost, Cumulative Expenditure, progress, dates, reporting agency, and line ministry remain separately typed. Identical organization names can occupy two exact published roles without being collapsed or inferred as a contractor.
• All 50 PAIMANA breadth markers remain representative because the Flash Report publishes no project coordinates or geometry. The selected OpenStreetMap entity, query, retrieval date, original coordinate, precision class, transformation, and limitation remain attributable per record.
• Two NHPC project pages add official locality context for Pakal Dul and Teesta VI. Their other cost, schedule, and status fields are not silently mixed with the PAIMANA snapshot and require a separate semantic review.
• Delhi Metro Phase V A and the Bhanupalli–Bilaspur–Beri line retain clearly labeled administrative or corridor-end navigation anchors where no source geometry was published. The markers are not presented as project extents or exact worksites.
• The report's six note-named inconsistent cumulative-expenditure project codes remain excluded, including the two that still appear in the table.
• Tender, award, signed contract, contractor, payment series, amendment, dispute, audit, package geometry, and final-cost evidence remain explicitly unlinked for the new basic records.
• Every active record remains an individual precision-aware feature. No lens contains clusters, count bubbles, state-count pills, or zoom-triggered aggregation.

Major-project breadth becomes genuinely national

Twenty-five source-backed major projects from MoSPI's May 2026 PAIMANA Flash Report join the default national layer across 16 states or union territories and 10 official infrastructure sectors. The map now opens with 40 Major projects, while all 124 toll-plaza facilities and local urban-service schemes remain preserved, searchable, and available through explicit layers.

Boundaries of this release

• Original Cost, Anticipated Cost, Cumulative Expenditure, physical progress, and original or revised completion dates keep their exact source meanings. Cumulative Expenditure is not relabeled as contractor payment, actual cost, or final cost.
• The source-named agency is displayed only as PAIMANA reporting agency. The line ministry or department remains separate, and no contractor, concessionaire, operator, selected bidder, or project company is inferred.
• PAIMANA publishes no project coordinates or geometry in the reviewed table. Each new marker is a cached, attributable OpenStreetMap facility, locality, or corridor representative and states that limitation directly.
• A report note says six project codes are omitted for inconsistent cumulative-expenditure data, but two still appear in the table. All six note-named codes are excluded from this batch, and the source contradiction remains documented.
• Tender, award, signed contract, contractor, payment series, amendment, dispute, audit, package geometry, and final-cost evidence remain explicitly unlinked for these basic records.
• No lens contains clustering, state-count pills, count bubbles, or automatic aggregation. Every record active in Major, Facilities, or All records remains an individual precision-aware map feature.

Major infrastructure leads, without hiding the evidence directory

The default national view now leads with 15 officially scoped central infrastructure projects, while all 124 toll-plaza facilities and local urban-service schemes remain preserved, searchable, and available through explicit Facilities or All records layers. Every active record renders as its own precision-aware mark at every supported zoom.

Boundaries of this release

• No project, package, tender, contract, payment, organization, dispute, audit, source, cost, schedule, or spatial record changed.
• National markers are intentionally small and crisp. Color follows the selected lens, while shape continues to carry spatial precision without relying on color.
• No lens contains a cluster, count bubble, state pill, or zoom-triggered aggregation. Coverage colors and sorts individual records by reviewed evidence depth.
• Toll plazas remain facilities unless an official parent-project relationship is linked. AMRUT locality records remain local schemes. The interface does not infer a parent or promote either class to Major.
• The primary canvas no longer displays an inactive Ask the map control. Search, four primary lenses, progressive lens access, compact filters, result count, and results form the browse first read.
• The current 139-record dataset has no identical coordinate pairs. Source-backed coordinates are not jittered. Future true collisions require zoom or explicit expansion.
• Infra retains direct MapLibre, OpenFreeMap Positron, official India geography, evidence-aware markers, and its single right workspace. No React wrapper migration or cross-repository map package is introduced.

Urban water and sanitation bring every state into the breadth lane

Thirty-five source-backed AMRUT 2.0 projects now join the public map, one for every state or union territory listed in the Ministry directory. The release takes the directory from 104 to 139 reviewed records while keeping each new project visibly basic until its implementing organization, tender, award, progress, schedule, payment, and deeper evidence are linked.

Boundaries of this release

• The AMRUT page publishes project title, type, literal Project Cost, milestone, and state and city or ULB scope. It does not publish a durable project id, project coordinate, publication date, implementing authority, contractor, tender, award, signed contract, progress, schedule, payment, amendment, dispute, or audit record.
• Project Cost remains the exact source label. It is not relabeled as original, sanctioned, tender, award, revised, expenditure, actual, or final cost, and no cost-change conclusion is calculated.
• A milestone such as Contract Awarded is a source-dated directory label. Without the underlying award record it does not identify the selected organization, award date, contract value, or current construction status.
• The official pages do not publish project geometry. A one-time, rate-limited, cached OpenStreetMap Nominatim match supplies a locality navigation point with attribution and provenance. Production makes no live geocoder request.
• The directory footer reports ₹1,97,191.85 crore, while its displayed state rows sum to ₹1,97,191.73 crore. The ₹0.12 crore source-display difference is retained instead of silently reconciled.
• Karnataka's empty, implausible city row is rejected because it lacks a city identity. Kargil and Bishnupur use reviewed town results rather than same-named district relations.
• The 35 records are a state-balanced first sample, not a claim that all 8,777 AMRUT projects have been reviewed. The NHAI breadth lane, six deep source ledgers, all prior releases, and every earlier source and spatial record remain intact.

Every mapped point gets district context

All 104 published points now have a source-backed mapped-district result. Readers can search and filter by district, see the district in hover or keyboard preview, and inspect the exact containment source and limitation in the Project Brief and Evidence view.

Boundaries of this release

• The filter is labeled Mapped district because it describes the district containing the displayed point, not necessarily the complete district scope of a project, corridor, package, or administrative-area record.
• Every district state and LGD state code agrees with the separate Release 22 state-containment result for the same unchanged point.
• A district result does not upgrade an editorial, locality, facility, corridor, or official-facility point to surveyed accuracy. Existing precision shapes and labels remain authoritative.
• Search, mapped-district filter, selected project, and camera remain shareable in the URL. The district query parameter is view state, not new spatial evidence.
• The analysis uses the 2026 Esri India service credited to Census of India, Survey of India, Local Government Directory, and Esri India. It is not a directly downloaded Survey of India polygon dataset.
• The public map remains at 104 records. This release improves geographic discovery and does not claim new project, contract, payment, cost, deadline, dispute, or audit evidence.

Six map conflicts get evidence-reviewed outcomes

The six Release 22 state conflicts no longer share one generic warning. Four now have an official-document-supported administrative state for discovery, while Mudipar and Moura remain visibly unresolved because the reviewed evidence cannot support a safe coordinate or state correction.

Boundaries of this release

• No original RajmargYatra state, coordinate, containment result, CRS, spatial type, or precision class was overwritten.
• The four reviewed administrative states agree with both a separate official document and the independent containment result. They improve filtering and state-count grouping but do not certify surveyed accuracy.
• Mudipar and Moura use a dashed outer marker shape, explicit high-attention wording, and a dedicated unresolved-location filter. They remain discoverable from both state signals.
• No authoritative replacement coordinate was located for Mudipar. The map does not substitute a locality, district centroid, corridor midpoint, or invented decimal point.
• Moura remains unresolved because the official Gazette itself combines Rupnagar district with Himachal Pradesh wording. The product does not choose a winner without a correction or newer authoritative record.
• A reviewed state difference or unresolved location is an evidence condition, not an allegation, audit finding, or conclusion about project quality, compliance, waste, misconduct, or corruption.
• The public map remains at 104 records. This release deepens spatial truth; it does not claim a new coverage count or add contract, payment, cost, deadline, dispute, or audit evidence.

Every marker gets an independent state check

All 104 unchanged marker coordinates now pass through a separate 2026 administrative point-in-polygon analysis. Ninety-eight agree with the project's declared source scope; six NHAI breadth records return a different state, and those conflicts are now prominent in hover, Project Brief, attention-filtered discovery, and the full Evidence view instead of being hidden by the source label.

Boundaries of this release

• The analysis uses the 2026 India Admin Boundaries state layer, credited to Census of India, Survey of India, Local Government Directory, and Esri India. It is a secondary dissemination service, not a directly downloaded Survey of India polygon file.
• The service receives each original EPSG:4326 point and returns only matching state attributes. No polygon geometry is downloaded, stored, simplified, republished, or bundled, respecting the source's no-offline-export boundary.
• A containment match checks the mapped point, not the full extent of a corridor or multi-state project, and it never upgrades an official facility or editorial point to surveyed accuracy.
• The six conflicts are source-review flags, not allegations, corrections, or proof that a project is misplaced. The original point, source state, independent result, and evidence limitation remain available together.
• The six conflicts are Mudipar: Chhattisgarh versus Rajasthan; Ghamroj and Jita Kherli Kiranj: Delhi versus Haryana; Moura: Himachal Pradesh versus Punjab; and Koyla and Sayed Majra: Uttarakhand versus Uttar Pradesh.
• No project, contract, payment, dispute, cost, deadline, or organization record was removed or relabeled. Release 21's 104-record directory and all earlier evidence remain intact.

The national map becomes a shareable evidence directory

All 104 records can now be narrowed by evidence depth, freshness flags, spatial precision, named organization coverage, and attention severity without weakening the underlying truth contract. Search, lens, filters, selected project, and map camera are written into the URL, so a reader can copy the exact investigation view from either browse or Project Brief.

Boundaries of this release

• The freshness filter surfaces deterministic review signals already supported by dated evidence. “No freshness flag” is not a guarantee that every source is current.
• Named source role means only that the reviewed record publishes a role and organization. It does not relabel a collection agency, concessionaire, operator, selected bidder, project company, or combined source role.
• Spatial filtering distinguishes exact points, official facility points, representative places, representative corridors, and conflicting official records. It does not upgrade any record’s precision.
• Camera coordinates in the URL describe the reader’s viewport, not project evidence. Project coordinates and their original provenance remain unchanged.
• At Release 21, administrative containment still relied on source state, highway, chainage, corridor, office context, and reviewed India bounds. Release 22 later adds the independent polygon check without rewriting this historical limitation.
• This release adds discovery behavior, not new project records or deeper contracts. Release 20’s 104 records and all prior evidence ledgers remain intact.

The map passes its first national coverage milestone

Eighty current NHAI toll-plaza records now enter through a reproducible breadth pipeline, taking the public map from 24 to 104 source-backed records. The batch spans 23 states and union territories while every new record remains visibly basic until construction, contract, payment, dispute, or audit evidence is linked.

Boundaries of this release

• New records publish only stable NHAI identity, active source-record state, supported corridor and office scope, official facility coordinates, highway and chainage, rate dates, user fee, and exact Collection Agency Name when present.
• Official facility coordinates are not presented as surveyed exact sites. The validator checks coordinate order, India bounds, the source-provided state, highway, chainage, and PIU or regional-office context.
• Active means only that the source record returned active when reviewed. It is not a construction, compliance, performance, traffic, revenue, or contract-quality judgment.
• A car user fee is not project cost, revenue, payment, award value, or final cost. Thirteen records with passed rate-revision dates stay explicitly stale rather than being silently treated as current.
• Collection Agency Name is preserved as that exact role. It is not relabeled as contractor, concessionaire, operator, selected bidder, or project company.
• Contact names, phone numbers, email addresses, nearby-service contacts, and embedded source documents are discarded before the reviewed snapshot is written.
• All 24 earlier records, six deep source ledgers, Project Brief behavior, map interaction, geography, and specialist evidence tabs remain intact.

Every project now opens with a complete first read

The selected-project Overview is now a calm, continuous Project Brief. It answers what the project is, where the marker comes from, what the latest reviewed record reports, how compatible costs or dates changed, who holds each sourced role, what contract evidence exists, why attention appears, and what remains unknown before asking a reader to open specialist tabs.

Boundaries of this release

• AI assisted the wording at build time, but structured official fields, deterministic comparisons, and linked sources remain the factual authority. No live AI request is added to hover or project opening.
• “No upward revision reported” describes the reviewed fields only. It is not a health, value, efficiency, or compliance judgment, and a revised value is never relabeled as actual or final cost.
• Organization names keep their exact sourced roles. Selected bidder, concessionaire, project company, operator, fee agency, payee, authority, and combined source roles are not silently converted into “contractor.”
• Attention signals explain the observed value or date and remain review prompts. The brief makes no supported conclusion about cause, blame, waste, misconduct, or corruption.
• Overview summarizes and links into Timeline, Money, Organizations, Packages, Contracts, and Evidence rather than duplicating each full dossier.
• All 24 records, Release 18 map behavior, and Releases 1 through 17 evidence remain intact. National coverage is still an incomplete sample.

India becomes an explorable map

The map now combines natural full-screen gestures, official state and union-territory orientation, a progressive city hierarchy, useful project previews, and spatial-precision-aware selection. It feels immediately explorable while every project placement and every geographic layer keeps its own evidence boundary.

Boundaries of this release

• The 1:16M state geometry is generalized orientation context, not cadastral, surveyed local, district, package, corridor, facility, or project-site geometry.
• City, town, and village labels use OpenMapTiles place points clipped to the official India outline. They are orientation labels, not evidence for a project position.
• Direct wheel, trackpad, drag, and pinch gestures are enabled on the full-screen map. Pitch and rotation remain constrained, and Reset to India remains available.
• A representative corridor still opens around its supported point because reviewed corridor geometry is not yet available. The camera does not imply a line or invented extent.
• Closing detail restores the prior browse camera unless the user deliberately changed the camera while detail was open. Hover and keyboard preview never move it.
• All 24 records and Release 17 evidence remain intact. National coverage is still an incomplete sample; Release 19 improves how each record is understood before breadth-lane expansion.

From solicitation to executed contract

Raha's 2022 fee-agency RFP now connects to a later NHAI executed-contract register row for the same plaza and procurement scope. The release adds the source-dated company, accepted Annual Remittance, contract period, and plaza code while preserving every missing document and current-status boundary.

Boundaries of this release

• The register names M B Consortium only as the company for a regular executed fee-agency contract from 9 December 2022 to 9 December 2023. The ended period does not establish the current agency or operator.
• The 2022 RFP still does not publish its bid-opening result, letter of award, or signed contract. The later register partially resolves the outcome without replacing those absent records.
• The ₹51.80 crore APC and ₹51.90 crore accepted Annual Remittance are compatible procurement fields. ₹1,245.88 crore project capital cost and ₹212.84 crore historic project-stretch revenue remain separate.
• M B Consortium in the executed register and M/s M.B. Consortium in NHTIS retain their exact strings and source roles. No construction role, transfer, legal-entity normalization, or current status is inferred.
• The register's PIU cell visibly reads “37 Nagaon,” while older and later official sources use PIU Nagaon and PIU Guwahati. The source difference remains visible.
• Actual collection, remittance receipts, amendments, performance, current rate and revenue, construction award and payments, revised or final cost, and audit findings remain unlinked.

An executed contract, not an inferred award

Pahammawlein adds Meghalaya's first reviewed fee-plaza trail and the product's first official executed fee-agency contract entry. The exact map point, NHTIS operating record, and NHAI contract register join through the same plaza name, km 13.550 chainage, NH-40 corridor, PIU, and regional office.

Boundaries of this release

• The register names M/s Skylark Infra Engineering Pvt. Ltd. only as the company for the executed regular fee-agency contract from 18 November 2022 to 18 November 2023. It is not relabeled as construction contractor or current operator.
• NHTIS separately names MB Consortium in its combined Concessionaire / OMT Contractor field. No transfer, corporate relationship, succession, or present-day role is inferred between the organizations.
• The ₹42.58 crore APC and ₹42.81 crore Annual Remittance are comparable procurement fields from one register row. ₹536 crore project capital cost and ₹24.41 crore historic project-stretch revenue remain separate.
• The contract period ended on 18 November 2023 and the NHTIS fee schedule was due for revision on 31 March 2026. Current agency, rate, revenue, remittance receipts, and performance require newer evidence.
• Jorbat-Shillong in the contract register and Jorabat-Shillong in the other NHAI records remain a visible source spelling difference.
• Signed contract text, tender and award file, payments, amendments, construction award and expenditure, revised or final cost, and audit findings remain unlinked.

Northeast evidence, without false equivalence

Raha (Saragaon) adds the first reviewed Assam fee-plaza trail: an exact NHAI point, the official NHTIS operating record, and a 2022 fee-agency RFP joined by the same plaza name and km 254.514 chainage. The public brief keeps procurement estimate, bidder remittance, project capital cost, historic toll revenue, and a stale fee schedule separate.

Boundaries of this release

• NHTIS reports commercial operation from 22 February 2017 and names M/s M.B. Consortium only in its combined Concessionaire / OMT Contractor role. It is not silently relabeled as the construction contractor or selected 2022 fee agency.
• The RFP publishes a one-year opportunity, 4 August bid deadline, 5 August scheduled opening, and a planned LOA window. It does not publish the opening result, award, selected agency, signed contract, accepted Annual Remittance, or actual collection.
• The ₹212.84 crore figure is cumulative toll revenue for the entire project stretch through 30 October 2024. It is not annual or current revenue, capital cost, an award, expenditure, or final cost.
• The accessible fee schedule was due for revision on 31 March 2026. Its ₹120 car fee is displayed as dated evidence, not claimed as the current 2026-27 rate.
• Nagao-Daboka-Udauli versus near-Nagaon to Guwahati corridor wording, NH-37 versus NH-27 numbering, and PIU Nagaon versus PIU Guwahati remain visible source differences.
• Underlying construction-package awards and payments, revised or final cost, current rate and revenue, amendments, and audit findings remain unlinked.

Four roles, one concession trail

Sadahalli now links its exact km 538 plaza and km 534.720-556.840 package to the original NHAI RFP and signed concession agreement. Selected bidder, project company, completed-projects organization, and current NHTIS organization remain four source-scoped identities rather than one guessed contractor history.

Boundaries of this release

• The ₹680 crore RFP estimate, agreement Total Project Cost ceiling input, and NHTIS capital cost share a number but not a definition. None is labeled actual or final cost.
• The ₹61.20 crore authority grant, ₹34 crore performance security, ₹13.60 crore bid security, ₹458.60 crore historic toll revenue, and ₹120 car fee remain separate money types.
• The agreement says Navayuga Engineering promoted the original project company. No reviewed source establishes the relationship to Navyuga Road Projects or the later Athaang NHTIS organization.
• The 30 April 2010 agreement is 81 days after the 8 February LOA, 36 days beyond the stated signing window. No extension or waiver is linked, so this is an evidence gap, not a delay or breach finding.
• The RFP's Bijapur-Hungund footer mismatch remains visible, while the project-specific title, package number, chainage, and signed agreement control the Sadahalli join.
• Grant disbursement, construction payments, amendments, final cost, current toll revenue, organization-transfer evidence, and audit findings remain unlinked.

One package, two plazas

Sardewadi joins Patas as a second mapped fee plaza on Pune–Solapur Package I. The product now carries one shared project-money record, one shared package award, a compatible pre-bid cost revision, and two explicitly conflicting official coordinate records without double counting or false precision.

Boundaries of this release

• The ₹1,371 crore NHTIS project capital cost and ₹378.70 crore project-stretch toll revenue are shared fields repeated on two plaza pages. Neither is doubled.
• The ₹922 crore to ₹1,110 crore change is a compatible bid-document revision before bid submission, not a later construction overrun, actual cost, expenditure, or final cost.
• The later ₹1,371 crore capital-cost field has a different source definition and is not appended to the bid-document series.
• Patas's two official coordinates are 4.92 km apart and Sardewadi's are 224.75 km apart. The later corridor-aligned IHMCL points are mapped as indicative facility points, while both conflicts remain visible.
• The ₹299 crore authority grant, 21-year concession, 910-day construction period, and ₹55.5 crore performance security belong to one shared package award, not separate plaza awards.
• Current toll revenue, plaza-level collection, final project cost, construction payments, amendments, audit findings, and the reason for the coordinate conflicts remain unlinked.

Four payments are not one contract

Tundla now connects four exact NHAI cheque rows from 2011 and 2012 to a separately bounded 2013 fee-agency RFP, without implying that the payee won the later procurement or that the reviewed rows form a complete contract.

Boundaries of this release

• ₹33,63,639 is only the sum of four reviewed rows, not a contract value, complete payment series, monthly amount, project cost, toll revenue, or final cost.
• The source spelling “M/s Cannon Mangement Security Services” remains exact; no silent organization-name correction or entity resolution is claimed.
• The repeated ₹14,58,887 amount under two different cheque references is not interpreted as either a contractual pattern or a duplicate error.
• The 2013 RFP does not identify a selected agency, award, filled Annual Remittance, actual collection, or relationship to the earlier payee.
• AE Tollway Limited remains in the current combined NHTIS role; the evidence does not connect it to Cannon or to the RFP outcome.

One plaza, three procurement baselines

Barsoni now connects three official NHAI fee-agency RFPs from 2014, 2019, and 2024 to one separately typed performance-security refund, while keeping every missing award result and actual collection figure visible.

Boundaries of this release

• Annual Potential Collection is NHAI's procurement estimate, not actual toll revenue, a bidder offer, award amount, authority receipt, project cost, or final cost.
• The ₹55,41,415 July 2014 entry is one cash performance-security refund to Sahakar Global, not construction expenditure, toll collection, or total payment.
• None of the three linked RFPs publishes its selected agency, bid-opening result, letter of award, signed contract, bidder payment, or actual collection series.
• The 2024 scope adds adjacent toilet-block upkeep, so the APC sequence is contextual procurement evidence rather than a pure like-for-like performance measure.
• One 2014 schedule page says West Bengal while the project-specific RFP sections, NHTIS, and official point agree on Bihar. The discrepancy remains visible.

The post-concession gap gets a tender trail

Laling now links IRCON's reported 21 April 2026 concession end to NHAI's 6 July 2026 solicitation for a user-fee agency covering the named Laling and Chandwad plazas.

Boundaries of this release

• The tender covers Laling and Chandwad together. Its scope and any future result are not attributed to Laling alone.
• Publishing a user-fee agency tender does not establish an award, who collected fees after 21 April, or who operates the corridor now.
• A user-fee agency is not relabeled as the construction contractor, concessionaire, project company, or corridor operator.
• Missing tender value, EMD, award amount, selected agency, and award date remain “not published,” never zero.
• The official active-tender list is dynamic. The exact tender ID is preserved for portal lookup while a stable direct detail or award-result link remains missing.

The historic claim gets a dated outcome

Laling's Pimpalgaon–Dhule trail now moves from NHAI's 2018 arbitration disclosure to two final Delhi High Court orders dated 16 April 2024, plus a separate IRCON operating-company record dated 31 March 2025.

Boundaries of this release

• Withdrawal after settlement is not presented as a win, loss, admission, liability finding, acceptance of the original demand, or evidence of misconduct.
• The Court's direction returning NHAI's deposit plus interest to NHAI is a court-deposit disposition, not a published settlement amount, payment, or project recovery.
• IRCON's operating status is dated 31 March 2025. Its reported 21 April 2026 concession end has passed, so current operation now requires a newer source.
• NHTIS, IRCON, and NHAI publish different concession dates, chainage limits, or rounded lengths. All remain visible instead of being silently reconciled.
• The settlement instrument, award and settlement amounts, full payment series, final project cost, amendments, and audit findings remain unlinked.

India returns to the center

The map now owns the majority of the screen. One adaptive workspace on the right alternates between discovery and a plain-language project brief, while the full official India outline, selected marker, filters, lens, results, and camera remain connected.

Boundaries of this release

• Only the 6 toll plazas use exact published coordinates. The other 15 markers are visibly encoded and described as representative facility, locality, or corridor points.
• A representative point is not project geometry, a package extent, a surveyed site, or proof that all work occurs at that point.
• “No upward revision reported” compares compatible fields in one source record. It is not a final-cost claim or a green verdict about project health.
• Organization labels preserve the role printed by the source. A concessionaire, operator, selected bidder, authority, and combined source role are not silently converted into contractor.

A second contract history, with allegations kept honest

Laling now connects the Pimpalgaon–Dhule package to a 2012 positive-grant payment and NHAI's 2018 disclosure of a ₹942.52 crore arbitration claim, without presenting either as project cost or carrying an old dispute status forward as current.

Boundaries of this release

• NHAI described the arbitration as pending on 16 November 2018. No later outcome is linked, so the product does not call it currently pending.
• One positive-grant payment is not the total grant, total disbursed, accepted contract value, project cost, or final cost.
• The ₹942.52 crore figure is the claimant's disclosed demand, not a tribunal award, payment, accepted liability, or verified loss.
• The full grant and payment series, amendments, original award, final cost, audit findings, and current dispute outcome remain unlinked.

From one payment to live balance works

Pottipadu now links its road package to an official 2019 payment entry and the government portal's current ₹21.59 crore balance-works tender, while keeping payment, tender value, project cost, and user fee separate.

Boundaries of this release

• A 2018 NHAI planning-status report prints Package II's exact km 1050.680–1076.480 limits, upgrading Pottipadu from medium to high join confidence.
• One IPC payment is not total expenditure, accepted contract value, final cost, or proof of completion.
• The ₹21.59 crore balance-works tender value is not an award amount, payment, cost overrun, or final project cost.
• “Under Risk and Cost of the EPC Contractor” is the official tender title. It is not converted into a finding about liability, cause, recovery, or misconduct.
• The earlier award, complete payment series, amendments, recovery outcome, and audit evidence remain unlinked.

Every toll gets a package trail

All six toll-plaza records now reach an official highway-package candidate, with five exact-chainage joins and one clearly labeled medium-confidence corridor join.

Boundaries of this release

• A package trail is a record-linkage claim, not proof of a complete contract, payment, or ownership history.
• Pottipadu is not upgraded to high confidence without a source printing the package's own exact endpoints.
• NHAI's combined contractor or concessionaire register field remains exactly that; it is not silently relabeled as awarded contractor.
• Original award evidence is still available only for Patas. Awards, amendments, payments, final costs, and audit findings for the other packages remain unlinked.

The first award trail

Patas now reaches the original NHAI letter of award and acceptance, connecting the physical toll plaza to the selected bidder and the financial and delivery terms that NHAI actually accepted.

Boundaries of this release

• The award evidence comes from the original NHAI concession volume containing the signed LOA and bidder acceptance.
• The ₹299 crore figure is a grant quoted by the selected bidder and payable by NHAI under the RFP. It is not labeled contract value.
• A 22 versus 24 December 2008 bid-date discrepancy inside the official volume remains visible.
• Bid evaluation, construction subcontracts, grant disbursement, amendments, payments, final cost, and audit findings remain unlinked.

Packages beneath the dots

Sadahalli and Patas now connect their fee-plaza records to official highway-package evidence, with organization roles and source differences kept visible.

Boundaries of this release

• A package join is labeled high confidence only when the official chainage, highway, package name, and plaza position align.
• NHAI's completed-projects register combines contractor and concessionaire in one column, so the product preserves that combined label.
• The Pune–Solapur Independent Engineer RFP cites a concession agreement; it is not presented as the original award record.
• Original award recommendations, works subcontracts, amendments, payments, and audit findings are still not joined.

Toll dots become highway records

Every toll-plaza marker now resolves to an officially published coordinate and carries the highway context and named operator available in NHAI records.

Boundaries of this release

• Coordinate evidence comes from an official 2017 NHAI plaza annexure and is linked separately from the NHTIS operating record.
• “Concessionaire / OMT contractor” is the combined source category; the product does not silently convert it into an awarded works contractor.
• Fee schedules may be stale when the source page’s revision-due date has passed.
• Contract documents, award value, amendments, and payments are still not joined.

The evidence-first map

A working full-viewport India map with the first source-backed national project layer and the system for inspecting evidence without leaving geography behind.

Boundaries of this release

• Project markers are an editorial sample, not complete nationwide coverage.
• Most toll and corridor coordinates are approximate locality or corridor placements.
• Tender packages, contracts, amendments, expenditure, and audit findings are not linked yet.
• Attention signals are arithmetic prompts, not allegations of corruption.