generated: '2026-08-13' method: probed status: published source: https://mcp.stackadapt.com/ docs: https://www.stackadapt.com/llms.txt server: name: stackadapt transport: http protocol: Streamable HTTP (JSON-RPC 2.0 over POST, MCP session via Mcp-Session-Id) url: https://mcp.stackadapt.com/ backing_api: https://api.stackadapt.com/graphql deployment: mode: remote endpoint: https://mcp.stackadapt.com/ install: null package: null auth: oauth verified: probed announcement: generally_available: '2026-04-21' source: https://www.businesswire.com/news/home/20260421122109/en/StackAdapt-Brings-Campaign-Intelligence-Into-Claude-and-Other-AI-Workflows-With-Launch-of-MCP-Server summary: >- StackAdapt announced general availability of its MCP Server on 2026-04-21, giving AI clients conversational access to campaign configuration, performance metrics and creative assets across CTV, display, native, audio, DOOH and programmatic linear TV. StackAdapt's own llms.txt describes it as the integration that extends Ivy(TM) into Claude and other AI workflows. authorization: model: OAuth 2.1 with RFC 9728 protected-resource discovery protected_resource_metadata: https://mcp.stackadapt.com/.well-known/oauth-protected-resource authorization_server: https://www.stackadapt.com/ authorization_server_metadata: https://www.stackadapt.com/.well-known/oauth-authorization-server dynamic_client_registration: https://www.stackadapt.com/oauth/register scopes: - graphql-public:read - graphql-public:write pkce: [S256, plain] grant_types: [authorization_code, client_credentials, refresh_token] challenge_observed: >- WWW-Authenticate: Bearer error="invalid_token", error_description="Missing Authorization header", resource_metadata="https://mcp.stackadapt.com/.well-known/oauth-protected-resource/" tools: status: gated count: null note: >- tools/list and initialize both return HTTP 401 invalid_token without a bearer token, so the live tool manifest and its inputSchemas require authenticated introspection. NO TOOL LIST IS RECORDED HERE. StackAdapt does not publish a tool catalogue in its llms.txt or on any crawlable page, and docs.stackadapt.com disallows all crawlers in robots.txt, so there is no public source to fall back to. Capabilities are described in prose only (campaign configuration, performance metrics, creative assets); those are not tools and are not recorded as such. x-evidence: - fetched: '2026-08-13' url: https://mcp.stackadapt.com/ method: GET http_status: 401 body: '{"error":"invalid_token","error_description":"Missing Authorization header"}' - fetched: '2026-08-13' url: https://mcp.stackadapt.com/ method: POST tools/list http_status: 401 - fetched: '2026-08-13' url: https://mcp.stackadapt.com/.well-known/oauth-protected-resource method: GET http_status: 200 content_type: application/json - fetched: '2026-08-13' url: https://mcp.stackadapt.com/mcp method: GET http_status: 404 note: The MCP endpoint is the host ROOT, not /mcp. Recorded so a later run does not re-guess. third_party_servers: - name: aymanshaikh/stackadapt-mcp url: https://glama.ai/mcp/servers/aymanshaikh/stackadapt-mcp official: false note: >- Community-built stdio MCP server wrapping the StackAdapt GraphQL API. NOT first-party and NOT counted as a StackAdapt agent surface. Recorded only so it is not mistaken for the official hosted server in a later pass.