Published dataset catalog

Every collection below is available through GET /api/v1/datasets/{dataset} and its /{record_id} detail operation. These named reference entries document the existing dynamic routes, so they also work in SDK list/get, CLI and MCP.

Agencies: use organizations with filters={"type":"agency"}. Companies: use organizations with filters={"type":"company"}. Programs: use programs. There is no separate companies or agencies dataset.

The current allowlist contains 97 collections. GET /datasets is authoritative for the deployed service. Each linked reference includes its complete allowed field schema; a database relation absent from this catalog is not an API promise.

Stable-key IDs/cursors can survive builds; positional IDs/cursors are build-bound and positional lists cap at 10,000 matches. Read pagination.

Award datasets

Budgets and funding

Evidence and publications

Organizations and agencies

Programs

Reference and SBIR data

Technology and supply chain

Unavailable collections

evidence-scope, entity-links, summary-parent-fiscal-years and summary-recipient-fiscal-years are currently paused because their query plans need further work. Do not assume a public database relation or an older schema audit is readable through the API. Use supported relationship and summary collections where their documented grain fits your task.

The published corrections dataset contains editorial correction records. It is different from /api/v1/corrections, which holds your account’s submitted proposals.