{
  "markdown": "# @pipeworx/open-contracting\n\nOpen Contracting MCP — international public procurement: government tenders and\ncontract awards published under the OCDS (Open Contracting Data Standard).\nMirrors bulk OCDS releases from 19 national and subnational publishers —\n838,000+ contracting processes — refreshed weekly from the OCP Data Registry.\n\nPart of [Pipeworx](https://pipeworx.io) — an MCP gateway connecting AI agents to 1476+ live data sources.\n\n## Tools\n\n- `oc_tender_search(query, country, status, category, min_value, days, limit)` —\n  free-text search over tender title, buyer and description. See\n  [Language matching](#language-matching) — the query works in English even\n  where the data is not.\n- `oc_recent(country, days, limit)` — newest releases first, the \"what just came\n  out\" view.\n- `oc_process_history(ocid)` — every release for one contracting process, in\n  order, showing the tender → award → contract progression.\n- `oc_coverage()` — per-country publisher, release and process counts, and\n  latest-release freshness. Call it first to see what is actually there.\n\n## Auth\n\nNone for the caller. The pack is `injectSupabase` — it reads the OCDS bulk\npublications over PostgREST, so a query resolves against one backing store\nrather than depending on 19 separate upstream APIs being up at once.\n\nKeep the wording here descriptive of WHAT is returned, not of where it is\nserved from: `pnpm check:hosting-claims` fails the build on published copy that\ntells callers we hold a copy of upstream data, and this file is published (it\ngenerates `docs/reference/open-contracting/`). That gate red-lined main for nine\nconsecutive pushes on 2026-08-23.\n\n## Language matching\n\n**Only 7 of the 19 publishers file their tender text in English.** The rest\npublish in their own language, and the tool description and `country` argument\nare both in English, so an English question used to hit a silent zero:\n`{\"query\":\"Ministry of Health\",\"country\":\"Dominican Republic\"}` returned 0 of\n40,740 processes, while `\"Ministerio de Salud\"` returned rows.\n\n| Language | Publishers |\n|---|---|\n| Spanish | Dominican Republic, Peru, Uruguay, Honduras, Guatemala, Mexico (Oaxaca), Argentina (Mendoza) |\n| Albanian | Albania, Kosovo |\n| Croatian | Croatia |\n| Italian | Italy (ANAC) |\n| Thai | Thailand (Bangkok) |\n| English | Kenya, Nigeria, Ghana, Zambia, Liberia, Rwanda, Tanzania |\n\n`oc_tender_search` now expands the query instead of translating it. The phrase\nyou passed is always searched; alongside it go its equivalents in the language\nof the `country` you filtered to, drawn from a synonym table of the buyer and\nsubject words these corpora actually use (ministry / ministerio / ministria /\nministarstvo / ministero / กระทรวง, and ~45 more). Every spelling rides in one\nPostgREST `or=(...)`, which is a single SQL predicate — the branches union, so\na row matching several is still returned once and there is no second round trip.\n\nThree properties worth knowing:\n\n- **It expands, never replaces.** These corpora are mixed, not consistently one\n  language, so rewriting the query one way would drop whatever is filed the\n  other way. A native-language query keeps working and can only gain rows.\n- **The lookup is bidirectional, which is also the accent fix.** A caller who\n  types `Ministerio de Educacion` gets the stored `MINISTERIO DE EDUCACIÓN`,\n  because the unaccented Spanish token resolves through the same table to the\n  accented stored form. That keeps every pattern fully literal, so the trigram\n  index still applies.\n- **It says what it searched.** The response carries `query_match.searched` (the\n  spellings) and `query_match.publisher_language`. A zero result returns a\n  `hint` naming the publisher's language, rather than a bare `count: 0` that\n  teaches the caller the data is absent when it is present.\n\n`*` inside a spelling is the wildcard between words, so `ministerio*salud`\nmatches `MINISTERIO DE SALUD PÚBLICA`. Word order is tried both ways for\ntwo-word phrases, since it flips between languages (\"road construction\" is\nfiled as \"construcción de carreteras\").\n\nFor an English publisher no vocabulary is substituted — only connector\ntolerance and word order, so `Ministry Health` and `health ministry` both reach\n`MINISTRY OF HEALTH`.\n\n## Data sources\n\n- OCP Data Registry (bulk OCDS downloads): https://data.open-contracting.org/\n- Open Contracting Data Standard: https://standard.open-contracting.org/\n- Publishers include Guatecompras (GT), OECE (PE), DGCP (DO), ARCE (UY), ONCAE\n  (HN), ANAC (IT), Narodne Novine (HR), PPRC (XK), PPC (AL), BMA (TH), and the\n  national procurement authorities of KE, NG, GH, ZM, LR, RW, TZ.\n\n## Quick Start\n\nAdd to your MCP client (Claude Desktop, Cursor, Windsurf, etc.):\n\n```json\n{\n  \"mcpServers\": {\n    \"open-contracting\": {\n      \"url\": \"https://gateway.pipeworx.io/open-contracting/mcp\"\n    }\n  }\n}\n```\n\n### What this endpoint actually serves\n\n`tools/list` at `https://gateway.pipeworx.io/open-contracting/mcp` returns the tools in the table\nabove **plus the shared Pipeworx meta-tools** — `ask_pipeworx`,\n`discover_tools`, `search_within`, `remember`/`recall` and the rest of the\ngateway-wide set. So the tool count you see is larger than this table: a\nsingle-pack endpoint currently lists roughly 30 shared tools alongside the\npack's own. The connection's `initialize` response states its exact scope, and\nis the authoritative answer for a given day.\n\nThis is deliberate, not multiplexing by accident. The meta-tools are what let a\nscoped connection answer a question this pack does not cover — via\n`ask_pipeworx`, which routes across the whole catalog — without you adding a\nsecond MCP server. There is currently no way to mount a pack endpoint without\nthem; if the extra schemas cost you more context than the routing is worth,\nconnect to the full gateway once rather than to several pack endpoints.\n\nOr connect to the full Pipeworx gateway to get every pack's tools listed\ndirectly, instead of just this one's:\n\n```json\n{\n  \"mcpServers\": {\n    \"pipeworx\": {\n      \"url\": \"https://gateway.pipeworx.io/mcp\"\n    }\n  }\n}\n```\n\nBoth URLs reach the same gateway and the same 1476+ data sources. The\nonly difference is which pack's tools are listed **directly**; `ask_pipeworx`\nreaches all of them from either one.\n\n## Using with ask_pipeworx\n\nInstead of calling tools directly, you can ask questions in plain English —\nthis works on the pack endpoint above as well as on the full gateway:\n\n```\nask_pipeworx({ question: \"your question about Open Contracting data\" })\n```\n\nThe gateway picks the right tool and fills the arguments automatically.\n\n## More\n\n- [Docs and guides](https://pipeworx.io/docs)\n- [pipeworx.io](https://pipeworx.io)\n\n## License\n\nMIT\n",
  "bytes": 6716,
  "sha": "6927db90d53f8fdcdc6c86b21481dfcf6fd79b6ddf1863e8292c47036cde8138",
  "repo_slug": "pipeworx-io/mcp-open-contracting",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_io_github_pipeworx_io_open_contracting_83e448ad/readme"
}