Public-data tools for AI agents. Official sources, cited answers.

AI agents are increasingly asked questions that deserve a real answer from a real record — is this company registered, has this drug been recalled, what did this company actually file. Tributary builds MCP servers that fetch those answers from the authoritative source and hand back a citation with every one.

• SEC Filings  — company filings, XBRL financial data and full-text filing search, from SEC EDGAR.
• Company Screening  — UK Companies House lookups (profile, directors, beneficial owners, charges) plus US, EU and UN sanctions screening.
• Drug & Product Safety  — FDA recalls, official drug labelling, adverse event reports and ClinicalTrials.gov studies.

Primary sources only. Everything comes from the issuing authority's own API — the SEC, Companies House, the FDA. No third-party aggregators sitting between you and the record.

Every answer carries its source. Each response includes a link to the official document, so an agent can show its working rather than asking you to take its word.

Limits are stated, not hidden. Each server's page says plainly what it does not cover. A sanctions screen returns a warning that a name match is a lead to verify, not a determination. Adverse event counts come with the caveat that voluntary reports do not establish causation. Where a confidently wrong answer would do real damage, the caveat travels with the data.

Maintained. These are monitored and fixed, not published and abandoned.

Pay per call, no subscription and no minimum. You are charged for the calls your agent actually makes.