Data catalogue
Find datasets, series, publications and extracts across the marketplace.
- Warehouse rows
- 6.17B
- Structured DB
- 1,711.35 GB
- Live data store
- 189.73 GB
- Data tables
- 1,03,091
- Publications & data products
- 1,749
About these figures
BigQuery reports 6,167,191,751 physical rows across 89 native base tables in 9 catalogue warehouses, measured on . Structured DB measures logical bytes across the same tables. The headline excludes 22 verified staging and backup tables with 365,935,268 rows and 57.49 GB of logical storage. Separate publisher editions can contain overlapping observations.
The earlier 40-MCP inventory reported 1,115,017,091 records across warehouses, files and upstream sources, including 811,051,329 upstream survey records from MoSPI NADA.
Live data store measures 189.73 GB of live catalogue objects across 39 Cloud Storage buckets, measured on .
UNCTAD ingestion files account for 153.97 GB of additional live storage. Retained soft-deleted objects account for 109.84 GB. Raw publisher archives and their corresponding BigQuery tables are measured in their respective storage layers.
Data tables counts 1,03,091 publisher tables held across the marketplace MCPs, measured in each MCP's own store on 2026-11-10: one table, sheet, dataflow, report section or API dataset in one edition counts once, the way data.gov.in counts resources. Raw copies, backups and archive duplicates are not counted twice. They belong to 1,749 publications and data products — six Economic Survey reports are one publication — which hold 10,664 editions, 8,62,701 series and 19,717 documents. India Intelligence and Survey chapter packs are built from these MCPs and are not added. Geo-Spatial can also reach 2,36,742 data.gov.in resources live; they are not held and not counted.
The search index includes 18,933 datasets (tables, survey studies and map layers) and 493 individually indexed series. Separate live and archived access routes can have their own entries. These figures apply across the catalogue, regardless of filters.
Reference & capabilities
74 results
finish insight article
Resume after research (e.g. host-mode): explanations → coverage → outline → draft → verify → render. follow_ups=True runs server-side follow-up briefs for gaps.
generate findings
Scan every series, score patterns by materiality × surprise × recency, and store the top ones as findings F1..Fn — each with the computation that produced it.
generate insight article
End to end, in one call: ingest → findings → Charter → select → artifact checks → plan → research → explanations → coverage → outline → draft → verify (+1 revision) → render. Takes 5-10 minutes in server mode — clients with short HTTP timeouts should use start_insight_article + get_article_job instead.
get article artifact
Read draft.md, verification.json, final.md, final.html or charts/<id>.json remotely. Final artifacts are withheld when verification has become stale.
get article job
Status of a start_insight_article job: status (running | done | failed), current stage, progress 0-1, and — when done — the result with the article Markdown. Long-polls: waits up to wait_seconds (max 150) for the job to finish or change stage before answering, so an agent can simply call this again after each reply instead of sleeping. Pass wait_seconds=0 for an immediate snapshot.
get dataset
Dataset state: stages run, counts, outputs, and the Spec.
get dataset evidence
Read the exact stored structured result behind a candidate's source_ref. Host-imported evidence is labelled host_supplied_result.
get dataset job
Status of a background build: status (running | done | failed), stage, progress 0-1; when done, the workbook summary, download link and validation. Long-polls up to wait_seconds (max 30) for a change; pass 0 for an immediate snapshot. A job left by a previous process is marked interrupted and can be restarted.
get research briefs
Host-orchestrated mode: briefs with a ready-to-use sub-agent prompt, the evidence item schema and capability hints, so you can spawn your own sub-agents with your own MCPs. Return their results via submit_research_result.
get retrieval tasks
Host mode: tasks with a ready agent prompt and the value schema, for your own agents and MCPs. Return their values with submit_values.
get session
Session state: stages run, counts, article files, and the Charter.
import dataset rows
Import a structured host result by column mapping, without retyping numbers. result: {columns: [...], rows: [[...]]} or {rows: [{...}]}. mapping: {entity_column|entity_constant, period_column, metrics: [{metric, column, unit}], basis, source_type, source_url, source_title, rows_path?}. Evidence is retained, every value is gated, identical retries are idempotent. Task remains open for more batches;
ingest data
Load data and documents into a session (creating one unless session_id is given).
insight readiness
Check provider, storage and source credential configuration without provider calls. Configuration checks do not prove that a key, model or upstream is reachable; use discover_sources to probe upstream MCP connections.
inspect dataset cell
Explain one cell: chosen source, all candidates/conflicts, override, formula dependencies, and persisted evidence ids. IDs and period must match the spec.
inspect insight state
Read a persisted workflow artifact: findings, charter, hypotheses, briefs, explanations, coverage, outline, profile, tables, analyses or artifact_checks.
link sheets
Candidate join keys across tables (shared time axis, overlapping dimension values).
list evidence
The evidence ledger. status: accepted | parked | all.
list sources
Registered sources (SQL, web search, fact stores, browsers, …) with capabilities and cost rung, plus the capability vocabulary tasks are routed by.
list sources
Registered research sources with capabilities and cost rung, plus the capability vocabulary briefs use.
Inventory snapshot 15 Nov 2026. Publisher dates and individual source checks can differ.