generated: '2026-08-26'
method: probed
source: https://docs.tabby.ai/mcp
provider: Tabby
providerId: tabby
name: Tabby Docs MCP Server
summary: Tabby ships an official hosted MCP server over its documentation. It is documented at https://docs.tabby.ai/introduction/ai-tools
  with copy-paste Cursor and Claude Code configs, and answers tools/list anonymously. Tabby states explicitly
  that the server "covers the documentation only. It does not call the Tabby API or move money" — so this
  is a retrieval surface over the integration guide and the OpenAPI spec, not an agent surface onto Checkout/Payments/Disputes.
deployment:
  mode: remote
  endpoint: https://docs.tabby.ai/mcp
  auth: none
  verified: probed
  probe_prior: (never probed)
  probe: live
  probe_why: live
  checked: '2026-09-11'
  source: claimed-backlog re-probe 2026-09-11
x-evidence:
  fetched: '2026-08-26'
  url: https://docs.tabby.ai/mcp
  method: POST
  request: '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'
  http_status: 200
  content_type: text/event-stream
  tools_returned: 3
transport:
  protocol: Streamable HTTP
  accept: application/json, text/event-stream
  sse_endpoint_probe:
    url: https://docs.tabby.ai/sse
    status: 405
client_config:
  cursor: "{\n  \"mcpServers\": {\n    \"tabby-docs\": {\n      \"url\": \"https://docs.tabby.ai/mcp\"\
    \n    }\n  }\n}\n"
  claude_code: claude mcp add --transport http tabby-docs https://docs.tabby.ai/mcp
docs: https://docs.tabby.ai/introduction/ai-tools
tools:
- name: search_tabby
  title: Search documentation
  description: Search across the Tabby knowledge base to find relevant information, code examples, API
    references, and guides.
  inputSchema:
    type: object
    required:
    - query
    properties:
      query:
        type: string
        description: Search query
      language:
        type: string
        description: Filter to specific language code (e.g., 'zh', 'es'). Defaults to 'en'
    additionalProperties: false
- name: query_docs_filesystem_tabby
  description: Run a read-only shell-like query against a virtualized, in-memory filesystem rooted at
    / that contains ONLY the Tabby documentation pages and OpenAPI specs. Nothing runs on a real machine.
  inputSchema:
    type: object
    required:
    - command
    properties:
      command:
        type: string
        description: A shell command to run against the virtualized documentation filesystem (e.g. `rg
          -il "keyword" /`, `tree / -L 2`, `head -80 /path/file.mdx`).
    additionalProperties: false
- name: submit_feedback
  title: Submit documentation feedback
  description: Report a problem with this documentation site so the docs team can fix it.
  inputSchema:
    type: object
    required:
    - path
    - feedback
    properties:
      path:
        type: string
        minLength: 1
        description: The documentation page path the feedback is about.
      feedback:
        type: string
        minLength: 1
        description: A clear description of the documentation issue or suggestion.
    additionalProperties: false
notes:
- The server is anonymous — no OAuth challenge was returned and no /.well-known/oauth-protected-resource
  or /.well-known/oauth-authorization-server document is served on docs.tabby.ai (both 404).
- No MCP server exists over the transaction API. Every money-moving operation still requires a direct
  REST call to api.tabby.ai / api.tabby.sa with a Bearer secret key.
- The A2A agent card at /.well-known/agent-card.json describes the same documentation surface.
related:
  crosswalk: mcp/tabby-tool-crosswalk.yml
  agent_card: a2a/tabby-a2a.yml