Docs

Base URL https://api.texasrrcapi.com. JSON over HTTPS. Independent of and not affiliated with the Railroad Commission of Texas; built from its public files.

Quick start

Get a key on the pricing page (Free needs no card). The key is shown once.

curl https://api.texasrrcapi.com/v1/production?from=202501&to=202512 \
  -H "X-API-Key: $RRC_KEY"

That returns monthly statewide oil, gas, condensate and casinghead volumes for 2025.

Authentication

Send the key in X-API-Key. The MCP endpoint also accepts Authorization: Bearer <key>. Keys are stored only as SHA-256 hashes. Rotate with POST /v1/account/rotate (the old key stops working); manage billing with POST /v1/account/portal, which returns a Stripe billing portal link.

Plans and limits

GET /v1/account returns your plan, limits and usage this month. Identical GET requests may be served from the edge cache for up to five minutes; those do not count against your quota.

IDs

The same entity has the same ID in every dataset. Raw RRC columns are kept exactly as published; IDs are added beside them.

Typed copies of text columns end in __num (exact integers or decimals) or __date. They exist only where every value converts exactly. Decimals are returned as strings.

Provenance

Every response carries meta.snapshot_id and, per dataset used, source_release_id and conformed_release_id. Rows carry source_sha256 and source_uri: the hash and location of the original RRC file. Cite the snapshot ID to make a result reproducible.

Data notes

• Well locations (well_locations): well_surface, well_bottom and well_path, one row per mapped point or path, with LAT27/LONG27 (NAD27, the shapefile datum) and LAT83/LONG83 (NAD83), raw text plus exact __num copies, and the geometry as WKT and WKB. About 363,000 surface points carry only a county code in API (no API number assigned), so they have no api10.
• Completions (completions): one table per form and child record (packet, w2, g1, w2_casing_data, w2_formation_data, ...), joined on TRACKING_NO and PACKET_ID. A packet can appear again in later daily files; take the row with the latest source_file date per PACKET_ID. Two packet columns the RRC manual does not document are kept as undocumented_59 and undocumented_60.

Endpoints

Example: an operator

curl https://api.texasrrcapi.com/v1/operators/257097 -H "X-API-Key: $RRC_KEY"

{"data": {"organization": {"operator_id": "257097", ...},
          "counts": {"drilling_permits.da_permit": 19549, "wellbore_ewa.og_wellbore_ewa": 15181,
                     "inspections.inspections": 165, "violations.violations": 13, ...},
          "recent_permits": {"rows": [...], "truncated": false}},
 "meta": {"snapshot_id": "...", "sources": {"p5_organizations": {"source_release_id": "..."}, ...}}}

MCP for agents

Streamable HTTP at https://api.texasrrcapi.com/mcp (Pro and up). Tools: catalog, get_operator, get_well, get_lease, get_permit, get_production, query_sql. Results match the REST API, provenance included.

Claude Code

claude mcp add --transport http texasrrcapi https://api.texasrrcapi.com/mcp \
  --header "Authorization: Bearer $RRC_KEY"

Cursor and other clients (mcp.json)

{"mcpServers": {"texasrrcapi": {"url": "https://api.texasrrcapi.com/mcp",
  "headers": {"Authorization": "Bearer YOUR_KEY"}}}}

SQL

DuckDB SQL over "<dataset>"."<table>", read-only: no writes, no file or network access, settings locked. 15-second timeout; up to 10,000 rows (limit). Precomputed rollups live in agg.production_state, agg.production_district, agg.production_operator and agg.production_field.

curl https://api.texasrrcapi.com/v1/query -H "X-API-Key: $RRC_KEY" -H "content-type: application/json" \
  -d '{"sql": "SELECT district, sum(LEASE_OIL_PROD_VOL__num) oil FROM \"pdq\".\"og_lease_cycle\" WHERE CYCLE_YEAR_MONTH__num BETWEEN 202501 AND 202512 GROUP BY 1 ORDER BY 2 DESC"}'

Bulk downloads

GET /v1/bulk (Business and up) lists every table of the current snapshot as zstd Parquet with row count, size, SHA-256 and a signed download link valid for up to one hour. Add ?dataset=pdq for one dataset. Check the SHA-256 after download.

Change feed

GET /v1/changes?since=2026-10-01 or ?since=<snapshot_id> lists each new snapshot: which datasets got a new source release and how each table's row count moved. Poll it, then re-pull only what changed. Changes are reported per dataset and table, not per row.

Status

api.texasrrcapi.com/status shows every dataset's upstream schedule, last change and row count. JSON at /v1/status, no key required.

Errors