Adobe Customer Journey Analytics
Run cross-channel reports and build segments in Adobe Customer Journey Analytics.
Federal procurement intelligence: contracts, opportunities, entities, protests and budgets from FPDS, SAM.gov and…
Opens Agent Studio, where connecting is one click. The connector URL below works in any MCP client.
Paste it into any MCP client. Setup docs
The Tango MCP server, from MakeGov, gives Claude or ChatGPT a research interface onto United States federal procurement data. It consolidates FPDS, SAM.gov, USAspending and Grants.gov into one queryable surface covering contracts, live solicitations, vendor entities, bid protests and agency budgets — the raw material of a bid decision.
Verified connector
Listed by Anthropic as a partner connector in its Connectors Directory.
Connection checked by Agentman on .
Anthropic states this reflects the level of review a connector received, not a security audit.
Tool names from Anthropic's directory listing. This server requires sign-in, so we could not read tool descriptions or parameter schemas.
total_obligated as market size, because include_summary profiles at most 250 matching records — described as a hard cap, not a ceiling a narrow filter escapes. When more records match, the response sets is_capped_sample.org_distribution as keyed on awarding_office, so a summary cannot break spending down by funding organisation even when you filtered on funding organisation. That is a subtle way to get a defensible-looking wrong answer.resolve returns ranked guesses, not lookups. Tango grades matches high, medium or low and states that a low match should be verified before use. An agent that reports the top candidate without the grade has converted a probability into a fact.tools/list, so no schemas and no safety annotations were readable. Tool names come from Anthropic's directory, corroborated and extended by Tango's docs. No tool here carries a safety annotation we have seen.prompts/list.It needs an account, but a free tier exists. Tango's plans page states every plan including Free provides complete access to contracts, entities, opportunities, grants, protests and budget data. Free is capped at 100 requests per day and 25 per minute. Paid tiers from Small upward raise those limits and unlock Tango-computed enrichments such as resolved corporate ownership and OCR'd solicitation attachments.
It depends on the dataset, and Tango documents each one. Opportunities and entities refresh from SAM.gov on a cadence Tango describes as typically 20 to 60 minutes. Contract data follows the ingest pipeline, with FPDS updating twice daily and USAspending inputs daily. Forecast cadence varies by agency feed. That per-dataset disclosure is unusually specific for this catalogue.
Because they are the same server under two names, and the wire evidence proves it rather than the marketing copy. The RFC 9728 descriptor at govcon.dev names `https://tango.makegov.com` as its authorization server, and the 401 response carries a content-security-policy permitting connections to that host. Tango's docs confirm govcon.dev is the original endpoint, still supported and not deprecated.
It converts a company or agency name into ranked candidate identifiers. Government contracting has a severe identifier problem — UEI, CAGE, legacy DUNS and countless name variants — so resolve runs a Bayesian resolver over name plus optional state, city and freeform context. Free tier returns three candidates; paid tiers return five with a `match_tier` quality grade Tango tells you to verify before using.
One is a dispatcher over historical records, the other is typed to live solicitations. Tango's docs describe `search` covering 13 record types including contracts, IDVs, subawards, protests and budget accounts, selected by a type parameter. `search_opportunities` is purpose-built for open SAM.gov solicitations and agency forecasts, filtering on response deadline, set-aside type and notice type.
No, and Tango says so plainly. Its documentation warns against reporting `total_obligated` as market size, because the statistical profile returned by `include_summary` is computed over at most 250 matching records. That is a hard cap. When more records match, the response flags `is_capped_sample`. For true full-population totals Tango directs you to different, purpose-built queries.
Yes. Anthropic's directory snapshot of 2026-08-16 names four tools. Tango's own MCP documentation, fetched 2026-08-19, describes five — the same four plus `fetch_api_docs`, which pulls curated reference material for a Tango resource. The docs also list seven static MCP resources the directory does not surface at all, covering set-aside codes and NAICS lookups.
documentation URL Anthropic's directory publishes; 301-redirects to https://docs.makegov.com/mcp/, fetched 2026-08-19) · retrieved 2026-08-19docs.makegov.com/robots.txt allows all user agents with no disallow rules and points at a sitemap of 70+ pages; no Content-Signal directives are present (fetched 2026-08-19). tango.makegov.com serves no robots.txt and no llms.txt (both 404). govcon.dev/robots.txt returns HTTP 403 from its object store — an infrastructure response, not a policy signal · retrieved 2026-08-19https://tango.makegov.com as the sole authorization server and read as the sole scope) · retrieved 2026-08-19initialize POST to the endpoint returns HTTP 401 with www-authenticate: Bearer resource_metadata="https://govcon.dev/.well-known/oauth-protected-resource" and a content-security-policy whose only permitted connect target is https://tango.makegov.com · retrieved 2026-08-19permissions published as "Read only", four tool names, no prompt names) · retrieved 2026-08-16Connect Tango once and your agents call these tools on their own: on a schedule, in a workflow, with nobody at the keyboard.