generated: '2026-08-14'
method: searched
status: published
source: https://www.explorium.ai/.well-known/mcp/server-card.json
docs: https://developers.explorium.ai/mcp-docs/agentsource-mcp
deployment:
  mode: both
  endpoint: https://mcp.explorium.ai/mcp
  install: docker run -i --rm -e API_ACCESS_TOKEN=<key> explorium/mcp-explorium
  package: https://github.com/explorium-ai/mcp-explorium
  auth: oauth
  verified: probed
  note: >-
    Explorium ships BOTH shapes. The hosted remote endpoint is
    https://mcp.explorium.ai/mcp over Streamable HTTP with OAuth - an agent can
    reach it today with no installation. A self-hostable stdio build is also
    published as the MIT-licensed Docker image in explorium-ai/mcp-explorium and
    the PyPI package explorium-mcp-server, which authenticates with the same
    api_key value passed as API_ACCESS_TOKEN. The stdio path is materially
    staler than the hosted one - explorium-mcp-server last published 0.2.11 on
    2025-05-13, and the mcp-explorium repo was last pushed 2026-03-03, while the
    hosted server card was current at probe time.
server:
  name: explorium
  display_name: Explorium Business Data Hub
  version: 1.0.0
  transport: streamable-http
  url: https://mcp.explorium.ai/mcp
  legacy_sse_url: https://mcp.explorium.ai/sse
  website: https://www.explorium.ai/mcp/
  playground: https://www.explorium.ai/mcp-playground/
  privacy_policy: https://www.explorium.ai/privacy-policy/
  provider: Explorium
authentication:
  type: oauth2
  required: true
  challenge: >-
    An unauthenticated POST to /mcp returns HTTP 401 with
    WWW-Authenticate: Bearer realm="OAuth", error="invalid_token" and a JSON
    body {"error":"invalid_token"}. Observed 2026-08-14.
  authorization_server: https://mcp.explorium.ai
  metadata: well-known/explorium-oauth-authorization-server.json
  flows: [authorization_code, refresh_token]
  pkce: [S256, plain]
  dynamic_client_registration: https://mcp.explorium.ai/register
  scopes_published: false
  self_hosted_auth: >-
    The Docker/stdio build takes the REST api_key as the API_ACCESS_TOKEN
    environment variable instead of running the OAuth flow.
introspection:
  tools_list: gated
  probe:
    url: https://mcp.explorium.ai/mcp
    method: 'POST {"jsonrpc":"2.0","id":1,"method":"tools/list"}'
    http_status: 401
    body: '{"error":"invalid_token","error_description":"Missing or invalid access token"}'
    fetched: '2026-08-14'
  note: >-
    Live tools/list and initialize both require a bearer token, so per-tool
    inputSchema could not be read anonymously. The tool names and descriptions
    below are taken verbatim from Explorium's own anonymous server card, not
    inferred. Input schemas require authenticated introspection.
tool_list_divergence:
  note: >-
    Explorium publishes TWO tool lists that do not agree, and this matters to an
    agent choosing which to code against. The MCP server card (the machine-
    readable one, saved to well-known/) names 14 tools built around a unified
    fetch-entities / fetch-entities-statistics pair plus a
    dataset/export/cost-estimation group. The documentation and the published
    Agent Skill both name 11 tools split by entity type
    (fetch-businesses / fetch-prospects, fetch-businesses-statistics /
    fetch-prospects-statistics) with no export, dataset, cost-estimate or
    pricing tools. The server card is treated as authoritative here because it
    is the document the server itself publishes for pre-connection discovery.
  server_card_tools: 14
  documented_tools: 11
  documented_source: https://developers.explorium.ai/mcp-docs/agentsource-mcp
tools:
- name: fetch-entities
  description: Find companies or prospects using firmographic, technographic, role and intent filters.
- name: match-business
  description: Resolve a company name and/or domain to an Explorium business ID in bulk.
- name: match-prospects
  description: Resolve a person (email, or full name plus company) to an Explorium prospect ID.
- name: enrich-business
  description: >-
    Add firmographics, technographics, funding, financials, workforce trends and
    related attributes to matched companies.
- name: enrich-prospects
  description: Add contact details and professional profiles to matched prospects.
- name: fetch-businesses-events
  description: >-
    Retrieve business event signals such as funding rounds, hiring changes, new
    products and acquisitions.
- name: fetch-prospects-events
  description: Retrieve prospect-level events such as role changes and company moves.
- name: fetch-entities-statistics
  description: Return aggregate statistics describing how a filtered audience is distributed.
- name: autocomplete
  description: Resolve free-text input to the standardized filter values the search API accepts.
- name: estimate-cost
  description: Estimate the credit cost of an operation before it is run.
- name: show-sample
  description: Preview sample rows from a result table before exporting.
- name: export-to-csv
  description: Export a prepared result table to CSV.
- name: get-dataset
  description: Retrieve a previously saved dataset.
- name: show-pricing-plans
  description: Show the available credit packages and their pricing.
clients:
  documented:
  - Claude Desktop
  - Cursor
  - VS Code
  - n8n
  - Zapier
  - LangGraph
  - OpenAI
  - Base44
  docs: https://developers.explorium.ai/integrations/mcp-integrations
crosswalk: mcp/explorium-tool-crosswalk.yml
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com