generated: '2026-07-21'
method: searched
source: https://docs.trust3ai.com/how-to-guides/mcp-configuration.html
status: published
server:
  name: trust3ai-mcp-server
  transport: http
  transport_detail: streamable_http
  url: https://api.na.trust3ai.com/mcp
  probed: '2026-07-21'
  probed_status: 401
  probed_note: Endpoint is live and returns 401 without a scoped API key, as documented.
authentication:
  type: apiKey
  header: x-trust3ai-key
  scopes:
  - MCP (read)
  notes: Scoped API key generated in the Trust3 console (Settings -> API Keys -> Generate scoped key).
    MCP (read) is currently the only MCP scope and grants read-only access to governance context (inventory,
    policies, trust scores). Keys carry an expiry of 7/30/90 days, 1 year, or custom.
docs: https://docs.trust3ai.com/how-to-guides/mcp-configuration.html
capabilities:
- Query the governed AI agent inventory from external MCP clients (Cursor, Claude Desktop, custom agents)
- Look up policies and policy posture
- Ask trust-score and governance questions outside the Trust3 UI
client_config_example:
  mcpServers:
    trust3ai-mcp-server:
      transport: streamable_http
      url: https://api.na.trust3ai.com/mcp
      headers:
        Accept: application/json, text/event-stream
        x-trust3ai-key: <api-key>
deployment:
  mode: remote
  endpoint: https://api.na.trust3ai.com/mcp
  verified: probed
  probe: gated
  checked: '2026-08-12'
  source: catalog MCP census