21 answers covering the most common ways AI agents query US regulatory and public data through AI Analytics. Each answer routes to a concrete endpoint or canonical URL you can hit directly.
Call /api/v1/sanctions/check?name=YOUR_NAME. It checks the OFAC Specially Designated Nationals list (19,680 entries + 30k aliases) with fuzzy name matching. Each match returns the OFAC UID, sanctions program, and a link to the canonical /sanction/{uid} page.
For one-shot KYC, use the unified screen — accepts name OR identifier:
?name=Wells+Fargo?ticker=PFE (resolves via entity_master)?cik=0000078003?npi=1972902351 (direct OIG NPI match)Returns a 0-100 risk score covering OFAC + SAM + OIG + DOJ + CFPB + courts + Fed enforcement + NHTSA + CISA.
Use /api/v1/debarments/check?name=NAME. It screens both the HHS-OIG List of Excluded Individuals/Entities (83,256 entries) and SAM.gov federal-contractor exclusions (106,669 entries).
Each match links to a canonical page: /oig-exclusion/{id} or /sam-debarment/{id}, with exclusion type, statutory basis (42 USC §1320a-7), and reinstatement status.
Use /api/v1/debarments/check?name=NAME or browse /api/v1/sam-exclusions/recent. Each debarred entity has a canonical page at /sam-debarment/{ext_id}.
/cve/{cve_id} returns a canonical detail page combining CISA Known Exploited Vulnerabilities + NIST NVD data. Example: /cve/CVE-2024-21413 shows severity, CVSS score, ransomware-use flag, CWE codes, and CISA remediation deadline.
Recent CVEs: /api/v1/cisa/kev/recent or /api/v1/nvd/recent?severity=CRITICAL. By vendor: /api/v1/cisa/kev/by-vendor/{vendor}.
Recent: /api/v1/fda/recalls/recent?product=drug (or food / device).
Per-recall detail: /recall/{recall_number} — Class I/II/III classification, recalling firm, distribution pattern. Append .md for Markdown or .json for schema.org NewsArticle.
Recent: /api/v1/nhtsa/recent — filter by ?park_it=1 for the most-serious "do not drive" recalls.
By manufacturer: /api/v1/nhtsa/by-manufacturer/{co}. Per-recall canonical detail: /nhtsa-recall/{campaign_number} with safety summary, consequence, remedy, and park-it / park-outside / OTA-update flags.
By company: /api/v1/courts/by-party/{company} returns all federal court dockets (CourtListener data) where the company appears as a party.
Per-case detail: /case/{docket_id}. Filter by Nature of Suit:
By company: /api/v1/cfpb/by-company/{company}. Each complaint has its own page at /complaint/{complaint_id} with the consumer narrative (when published), issue/sub-issue classification, and company response.
NIH RePORTER: /grant/nih/{appl_id} — amount, recipient, PI, activity code (R01/R21/SBIR/etc.), abstract.
NSF: /grant/nsf/{award_id}.
Browse: /api/v1/nih/recent, /api/v1/nsf/recent. By organization: /api/v1/nih/by-org/{org}.
Recent: /api/v1/grants-gov/recent. Per-opportunity: /opportunity/{opp_id} — opportunity number, issuing agency, open/close dates, status, and CFDA codes.
Per-trial: /clinical/{nct_id} — phase, status, conditions, interventions, sponsor + collaborators, and the why_stopped reason if terminated.
Recent: /api/v1/clinical/recent?phase=PHASE3 for late-stage drug pipelines.
Per-facility: /facility/{frs_id} — owner organization, regulatory programs (CAA/CWA/RCRA/EPCRA/SDWA), compliance status, GPS coordinates.
Browse by owner: /api/v1/epa/facilities?owner={company}.
By client (the company being represented): /api/v1/lobbying/by-client/{company}. Per-filing: /lobbying/{filing_uuid} with registrant, quarterly income/expenses, and itemized activities keyed by issue code (HCR, ENG, TAX, etc.).
/api/v1/cross-vertical/active returns entities active in 2+ regulators simultaneously. Identifies multi-vertical regulatory exposure that single-regulator APIs miss.
/entity/{ticker-or-cik-or-uei} returns a unified cross-vertical timeline merging:
SEC filings, insider trades, Form 144, 13D/G, FDA recalls, drug labels, federal contracts, lobbying filings, clinical trials, court dockets, CFPB complaints, NIH grants, NHTSA recalls, DOJ press releases, Federal Reserve enforcement actions, EPA facility inspections, USPTO patents.
Each event hyperlinks to its canonical record page (the new ones we just shipped). JSON: /api/v1/entity/{key}/timeline. Material events only (importance ≥ 60): /api/v1/entity/{key}/material-events.
/today is an on-request cross-vertical digest. The default source-date window is 7 UTC calendar dates ending on the generation date; set ?days=1 through ?days=30.
The response publishes its UTC window and generation time. Active weather and forward-looking Form 144 sale dates are documented exceptions. Source freshness varies, so check /api/v1/freshness and /coverage before treating an empty section as proof that no event occurred.
Available formats: /today (HTML) · /today.md (Markdown for chat) · /today.json (JSON).
AI Analytics combines US federal sources with non-federal, self-regulatory, and third-party sources. The 229 datasets currently include:
SEC EDGAR · openFDA · OFAC Treasury · CFPB · NIST NVD · CISA KEV · NHTSA · DOJ · NIH RePORTER · NSF · ClinicalTrials.gov · EPA ECHO · USAspending · Federal Register · FINRA · CourtListener · OpenSanctions.
Each response should be read with its primary-source and source-terms metadata. Keyless access does not alter those terms.
See the full catalog at /datasets/. Live row counts + last-refresh timestamps at /coverage.
Live row counts + last-refresh timestamps for every dataset: /coverage. Machine-readable: /coverage.md, /coverage with Accept: application/json.
Source-level status: /api/v1/freshness. Cadence varies by source and recorded timestamps can lag. /today applies a real UTC source-date window to indexed records, but it does not prove that every upstream source refreshed inside that window.
There is no blanket license for every source. Works authored by US federal employees as part of official duties are generally public domain under 17 USC §105; that rule does not automatically cover every item on a federal site. Non-federal, SRO, and third-party material retains its source terms.
AI Analytics-authored metadata and normalization are CC0 1.0 only where expressly stated.
Every endpoint returns provenance fields such as primary_source, license, terms_of_use, and generated_at where applicable.
Suggested citation format:
AI Analytics, [dataset or record], retrieved [date] from [endpoint]; primary source: [source URL]; source terms: [license or terms URL].
Full license details at /license.
Yes. /mcp is a JSON-RPC over Streamable HTTP MCP server exposing 73 tools for cross-vertical regulatory lookups. No auth, no rate limit beyond ordinary anti-abuse.
Discovery:
/.well-known/mcp/server-card.json — SEP-1649 server card/server.json — registry.modelcontextprotocol.io manifestOpenAPI 3.1 spec: /openapi.json. No API key, no rate limit beyond ordinary Cloudflare anti-abuse.
Content negotiation on most endpoints:
Accept: text/html or browser → HTMLAccept: text/markdown or .md suffix → MarkdownAccept: application/ld+json or .json suffix on record pages → schema.org JSON-LD