Repository avatar
Search & Data Extraction
v0.1.0
active

gleif-lei-mcp

io.github.santex85/gleif-lei-mcp

Company lookup in the official GLEIF LEI registry: search, LEI records, ownership, validation.

Documentation

GLEIF LEI MCP server

MCP server for the Global LEI Index. It looks up legal entities (companies, funds, branches and government bodies) by name, fetches full LEI records, maps corporate ownership (parents and children) and validates LEI codes. No API key is needed.

It is ported from the gleif-lei-lookup Apify actor. Name matching and scoring, record flattening and parent lookups are the same, but there is no Apify dependency.

Tools

ToolWhat it does
search_companiesFuzzy name search in any language or script, with an optional country, ranked by a matchScore from 0 to 100. Legal-form words such as Ltd or GmbH are ignored.
get_leiFull record for one LEI: legal and HQ address, legal form, register ID, status, renewal date, BICs, and the direct and ultimate parent.
company_ownershipDirect and ultimate parent (or the official reason none is reported), plus direct and ultimate children with totals.
validate_leiFormat and ISO 17442 MOD 97-10 checksum check, then whether the LEI exists, its status and whether it is current.
list_entitiesPaged list of entities by country, entity status, registration status and category.

Example calls:

// search_companies
{"name": "Siam Commercial Bank", "limit": 3}
{"name": "Toyota Motor", "country": "JP"}
// get_lei
{"lei": "54930068OELHUBHH1U03", "include_ownership": true}
// company_ownership
{"lei": "5493006W3QUS5LMH6R84", "max_children": 10}
// validate_lei
{"lei": "5493006W3QUS5LMH6R84"}
// list_entities
{"country": "TH", "entity_status": "ACTIVE", "registration_status": "ISSUED", "category": "FUND", "limit": 25, "page": 1}

Sample search_companies result:

{"query": "Siam Commercial Bank", "count": 2, "results": [
  {"lei": "54930068OELHUBHH1U03", "legalName": "บริษัท ธนาคารไทยพาณิชย์ จำกัด (มหาชน)",
   "legalNameEn": "THE SIAM COMMERCIAL BANK PUBLIC COMPANY LIMITED", "country": "TH", "city": "BANGKOK",
   "status": "ACTIVE", "registrationStatus": "ISSUED", "category": "GENERAL",
   "registrationAuthorityEntityId": "0107536000102", "matchScore": 100.0}, "..."]}

Run

python3 -m venv .venv && .venv/bin/pip install -r requirements.txt
PORT=8094 .venv/bin/python src/server.py                 # streamable-http on 0.0.0.0:8094/mcp
MCP_TRANSPORT=stdio .venv/bin/python src/server.py       # stdio

Environment variables: HOST and PORT (default 0.0.0.0:8080), MCP_TRANSPORT (default streamable-http) and GLEIF_USER_AGENT.

Limits

  • GLEIF allows 60 requests per minute. The client keeps at least 1.1 s between requests across the whole process and backs off on 429 and 5xx responses: at most 3 tries, with waits capped at 15 s and a 20 s HTTP timeout.
  • search_companies returns at most 25 results. It takes 2 to 5 requests, so about 2 to 5 s.
  • get_lei with ownership takes 3 to 6 requests, about 4 to 7 s.
  • company_ownership takes about 6 to 9 requests, about 7 to 10 s. It returns at most 100 children per list, and total gives the full count.
  • list_entities returns at most 100 results per page.
  • Sole-proprietor LEIs describe natural persons. They are left out of searches and lists, and get_lei reduces them to their status.

Data source and licence

  • Data comes live from the public GLEIF API, which is built on the GLEIF Golden Copy. LEI data is published free of charge under CC0 1.0 (see GLEIF's LEI Data Terms of Use). Source: Global Legal Entity Identifier Foundation, https://www.gleif.org.
  • This server is not affiliated with or endorsed by GLEIF or any LEI issuer (LOU). Match scores and the flattened format are this project's own work.
  • Ownership data is GLEIF Level 2 accounting consolidation ("who owns whom"). It contains no ownership percentages and does not cover natural-person owners. When an entity reports no parent, the exceptionReason field says why, for example NATURAL_PERSONS, NON_CONSOLIDATING or NO_KNOWN_PERSON.
  • Fuzzy matching is statistical. Before making compliance decisions, confirm a match against the register number and address.