generated: '2026-08-17'
method: searched
status: published
source: https://docs.pyannote.ai/mcp
deployment:
  mode: remote
  endpoint: https://docs.pyannote.ai/mcp
  auth: none
  verified: probed
  note: Verified by POSTing initialize and tools/list anonymously on 2026-08-17; both returned HTTP 200
    with a text/event-stream JSON-RPC response. No API key, OAuth challenge or session header was required.
    There is no stdio/npx package — nothing to install.
  probe_prior: (never probed)
  probe: live
  probe_why: live
  checked: '2026-09-11'
  source: claimed-backlog re-probe 2026-09-11
server:
  name: pyannoteAI
  version: 1.0.0
  transport: http
  protocol_version: '2025-11-25'
  url: https://docs.pyannote.ai/mcp
  vendor_platform: Mintlify
  instructions: '"This Model Context Protocol server provides search and retrieval tools for the pyannoteAI
    site. Use it to answer questions from public site content. ... Apart from the submit_feedback tool,
    the server is read-only and scoped to pyannoteAI; it does not otherwise perform actions, mutate state,
    or access anything beyond the published site content and these resources."'
scope:
  kind: documentation
  note: IMPORTANT — this is a DOCUMENTATION-retrieval MCP server, not an API-operation server. Its three
    tools search and read the published docs; none of them call the pyannoteAI REST API. An agent can
    use it to learn how to diarize audio, but cannot use it to actually submit a diarization job. The
    ten REST operations in openapi/ have no MCP tool. See mcp/pyannoteai-tool-crosswalk.yml.
capabilities:
  tools:
    listChanged: true
  resources:
    listChanged: true
tools:
- name: search_pyannote_ai
  title: Search documentation
  description: Search across the pyannoteAI knowledge base to find relevant information, code examples,
    API references, and guides.
  input_schema:
    type: object
    required:
    - query
    properties:
      query:
        type: string
        description: Search query
  annotations:
    readOnlyHint: true
    destructiveHint: false
    idempotentHint: true
    openWorldHint: false
- name: query_docs_filesystem_pyannote_ai
  description: Run a read-only shell-like query (rg, grep, find, tree, ls, cat, head, tail, jq, ...) against
    a virtualized in-memory filesystem containing only the pyannoteAI documentation pages and OpenAPI
    specs. Stateless per call; no writes, no network.
  input_schema:
    type: object
    required:
    - command
  annotations:
    readOnlyHint: true
- name: submit_feedback
  title: Submit documentation feedback
  description: Report an incorrect, outdated, confusing or incomplete documentation page to the docs team.
  input_schema:
    type: object
    required:
    - path
    - feedback
  annotations:
    readOnlyHint: false
  note: The only non-read-only tool on the server; it writes documentation feedback, not API state.
resources:
- uri: mintlify://skills/pyannote
  name: pyannote
  mime_type: text/markdown
  description: Use when building speaker diarization and identification features into voice applications,
    meeting transcription systems, call center analytics, live captioning, or any audio processing workflow
    that needs to determine who spoke when and identify specific speakers using voiceprints.
  note: Same Agent Skill advertised by the A2A agent card and served at https://docs.pyannote.ai/.well-known/agent-skills/pyannote/skill.md
    — harvested to skills/pyannoteai-pyannote-published.md.
x-evidence:
  fetched: '2026-08-17'
  url: https://docs.pyannote.ai/mcp
  method: POST
  requests:
  - body: '{"jsonrpc":"2.0","id":1,"method":"initialize",...}'
    http_status: 200
  - body: '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'
    http_status: 200
    tools_returned: 3
  - body: '{"jsonrpc":"2.0","id":2,"method":"resources/list"}'
    http_status: 200
    resources_returned: 1
  content_type: text/event-stream
  hosts_probed_and_missed:
  - url: https://mcp.pyannote.ai/mcp
    status: 0
    note: DNS does not resolve
  - url: https://api.pyannote.ai/mcp
    status: 404
  - url: https://www.pyannote.ai/mcp
    status: 404