generated: '2026-08-02'
method: searched
status: published
source: https://docs.aigateway.cequence.ai/docs/remote-mcp-servers/cequence-ai-gateway
llms_source: https://docs.aigateway.cequence.ai/llms.txt
server:
  name: Cequence AI Gateway MCP
  vendor_first_party: true
  description: >-
    First-party, Cequence-hosted MCP server to manage the AI Gateway from any agent — create MCP
    servers, agent personas, skills, and DLP policies. Provisioned automatically for every tenant.
  transport: streamable-http
  url: null
  url_note: >-
    Per the docs, "Server URL: Shown on the server's details page in the UI - copy it from there to
    connect your client." The endpoint is tenant-specific, so there is no single public MCP URL to
    probe and tools/list could not be called anonymously.
  provisioning: automatic per tenant
auth:
  methods: [oauth2]
  oauth2:
    description: >-
      OAuth 2.0 / OIDC — browser-based authorization-code flow against the AI Gateway account
      (identity provided by Descope), standard public-client PKCE flow.
    grant_types: [authorization_code]
    pkce: true
    identity_provider: Descope
  discovery:
    oauth_authorization_server: not published anonymously (SPA catch-all returns HTML 200 for every /.well-known/* path on aigateway.cequence.ai)
    oauth_protected_resource: not published anonymously (same SPA catch-all)
tool_count: 25
tools_source: >-
  Tool names and descriptions captured verbatim from the provider's published documentation page.
  inputSchema was NOT captured — the live endpoint is tenant-scoped and OAuth-gated, so authenticated
  MCP introspection (tools/list) is required to obtain real parameter schemas. Do not assume schemas.
tools:
- {name: list_third_party_apps, category: catalog, description: "Browse the Cequence-curated catalog of apps you can build an MCP server from"}
- {name: get_third_party_app, category: catalog, description: "Fetch one catalog app's endpoints (API) or tools (MCP)"}
- {name: list_api_specs, category: specs, description: "List the tenant's uploaded custom API specs"}
- {name: get_api_spec, category: specs, description: "Fetch one API spec's endpoints"}
- {name: list_mcp_specs, category: specs, description: "List the tenant's saved remote-MCP specs"}
- {name: get_mcp_spec, category: specs, description: "Fetch one MCP spec's tool inventory"}
- {name: list_mcp_servers, category: servers, description: "Discover deployed MCP servers (gateways) in the tenant"}
- {name: list_tools_for_mcp_server, category: servers, description: "List a single MCP server's tool definitions"}
- {name: list_apis, category: api_registry, description: "List registered REST APIs (the API Registry)"}
- {name: get_api_registry, category: api_registry, description: "Fetch a registered API's operations"}
- {name: list_pools, category: deployment, description: "List private deployment pools"}
- {name: list_teams, category: access_control, description: "List teams (SSO-mapped user groups)"}
- {name: list_skills, category: skills, description: "List registered skills"}
- {name: get_skill, category: skills, description: "Fetch one skill's full content"}
- {name: list_sdp_categories, category: dlp, description: "List standard sensitive-data-protection categories"}
- {name: list_dlp_policies, category: dlp, description: "List DLP policies (optionally scoped to a persona)"}
- {name: list_agent_personas, category: personas, description: "List agent personas"}
- {name: get_agent_persona, category: personas, description: "Fetch one persona, including its connect URL"}
- {name: query_tool_activity, category: observability, description: "Observability: logs, summary, facets, trends, log detail, and findings"}
- {name: recommend_dlp_policy, category: dlp, description: "Recommend a DLP policy for a persona (deterministic)"}
- {name: create_mcp_server, category: servers, description: "Create an MCP server from an API or remote-MCP spec"}
- {name: create_agent_persona, category: personas, description: "Create an agent persona (supports dryRun validation)"}
- {name: update_agent_persona, category: personas, description: "Apply partial edits to an existing persona"}
- {name: create_skill, category: skills, description: "Create a reusable skill in the Skill Registry"}
- {name: attach_dlp_policy, category: dlp, description: "Create and attach a DLP policy to an existing persona"}
clients_documented:
- {client: Claude Desktop, docs: https://docs.aigateway.cequence.ai/docs/client-config/claude-desktop}
- {client: Claude Code, docs: https://docs.aigateway.cequence.ai/docs/client-config/claude-code}
- {client: Cursor IDE, docs: https://docs.aigateway.cequence.ai/docs/client-config/cursor}
- {client: VS Code, docs: https://docs.aigateway.cequence.ai/docs/client-config/vscode}
- {client: Windsurf, docs: https://docs.aigateway.cequence.ai/docs/client-config/windsurf}
frameworks_documented:
- {name: LangGraph, docs: https://docs.aigateway.cequence.ai/docs/ai-libraries/langgraph}
- {name: CrewAI, docs: https://docs.aigateway.cequence.ai/docs/ai-libraries/crewai}
setup_cli:
  package: '@cequenceai/mcp-cli'
  see: cli/cequence-security-cli.yml
registry_listings:
  modelcontextprotocol_registry:
    listed: false
    probe: https://registry.modelcontextprotocol.io/v0/servers?search=cequence
    result: 0 servers
notes: >-
  Cequence AI Gateway is itself an MCP *production* platform — it converts customer REST APIs and
  third-party remote MCP servers into governed MCP endpoints. This artifact records only the
  first-party MCP server Cequence operates for managing the gateway. The 60+ third-party remote MCP
  servers documented under /docs/remote-mcp-servers/ belong to other vendors and are not Cequence
  APIs.
x-evidence:
  fetched: '2026-08-02'
  urls:
  - url: https://docs.aigateway.cequence.ai/docs/remote-mcp-servers/cequence-ai-gateway
    http_status: 200
  - url: https://docs.aigateway.cequence.ai/llms.txt
    http_status: 200
    content_type: text/plain