generated: '2026-08-13' method: searched status: published source: https://docs.tryprofound.com/mcp/overview deployment: mode: remote endpoint: https://mcp.tryprofound.com/mcp auth: oauth verified: probed note: 'Remote-only. Profound ships no stdio package — there is no npx/pip MCP server to install; clients connect directly to the hosted URL. Confirmed by POSTing tools/list to the endpoint: it answered HTTP 401 with a WWW-Authenticate Bearer challenge and a resource_metadata pointer, which proves a live MCP server is listening and that the tool schemas are auth-gated.' probe_prior: (never probed) probe: gated probe_why: RFC 9728 challenge on the MCP path only checked: '2026-09-11' source: claimed-backlog re-probe 2026-09-11 server: name: profound_mcp transport: http url: https://mcp.tryprofound.com/mcp docs: https://docs.tryprofound.com/mcp/overview authentication: primary: oauth2.1 docs: https://docs.tryprofound.com/mcp/authentication authorization_server: https://auth.tryprofound.com protected_resource_metadata: https://mcp.tryprofound.com/.well-known/oauth-protected-resource authorization_server_metadata: https://mcp.tryprofound.com/.well-known/oauth-authorization-server dynamic_client_registration: true pkce: - S256 scopes_supported: - email - offline_access - openid - profile alternative: method: bearer-api-key note: Long-lived Bearer token using a Profound API key, for service accounts and internal tools that cannot complete a per-user OAuth flow. Requires the Enterprise plan and API access granted by Profound support. behavior: analytics_tools_read_only: true hints: readOnlyHint: true idempotentHint: true note: 'Read-only hints apply to the analytics/report tools. The Agents tools do create, update and publish agent definitions and start runs; they default to preview: true so a client reviews a plan before it is applied.' tools: discovery: - name: whoami description: Confirm the authenticated principal, organizations, regions, and entitlements - name: list_organizations description: List organizations the user can access - name: list_regions description: List geographic regions, optionally scoped to an organization - name: list_models description: List tracked AI models such as ChatGPT, Perplexity, Google AI Overviews, and Gemini - name: list_categories description: List tracked categories for an organization - name: list_domains description: List tracked domains for an organization - name: list_tags description: List tags in a category - name: list_topics description: List topics in a category - name: list_prompts description: List configured prompts in a category, with optional status, tag, and topic filters brand_visibility_reports: - name: get_visibility_report description: How visible a brand is in AI answers, by model, topic, region, prompt, tag or persona - name: get_sentiment_report description: What sentiment AI answers express about a brand or competitor, and which themes are positive or negative - name: get_citations_report description: Which domains, pages or URLs AI engines cite for a category - name: get_prompt_answers description: The raw AI answers observed behind the metrics - name: get_factcheck_report description: FactCheck scores for a category over a date range - name: get_factcheck_claims description: Inaccurate claims identified in a category over a date range traffic_reports: - name: get_referrals_report description: How many visits a domain received from AI engines, and which referrers drove them - name: get_bots_report description: Which AI crawlers are visiting a domain, and how often agents: - name: list_agents description: Browse Agents in the organization - name: get_agent description: Read an Agent and its input_schema before running it - name: run_agent description: Start a run with inputs matching the Agent's input_schema - name: get_agent_run description: Poll a run by agent_id and run_id until it reaches a terminal state - name: list_agent_definition_templates description: List pre-built Agent templates - name: list_agent_node_types description: List workflow node types available for building an Agent graph - name: get_agent_node_schema description: Get the schema for one Agent node type - name: start_agent_build_session description: Open a traced build session from a plain-language intent - name: get_agent_definition description: Read an Agent definition draft - name: create_agent_definition description: Create an Agent definition draft (preview by default) - name: update_agent_definition description: Update an Agent definition draft (preview by default) - name: validate_agent_definition description: Pre-check an Agent definition for structural issues - name: publish_agent_definition description: Publish an Agent definition to make it live (preview by default) resources: - uri: file:///profound/glossary audience: assistant description: Compact index of Profound-specific terms to load once per session - uri: file:///profound/glossary/full audience: user description: Full glossary with definitions and examples - uri: file:///profound/glossary/{term} audience: assistant description: Full definition of one glossary term, addressed by slug clients_documented: - Claude Code - Claude Desktop - ChatGPT - Cursor - Gemini CLI - Microsoft Copilot Studio - OpenCode - VS Code - Windsurf - WRITER AI Studio x-evidence: fetched: '2026-08-13' probes: - url: https://mcp.tryprofound.com/mcp method: POST tools/list http_status: 401 note: WWW-Authenticate Bearer error="invalid_token"; live server, schemas auth-gated - url: https://mcp.tryprofound.com/.well-known/oauth-protected-resource http_status: 200 - url: https://mcp.tryprofound.com/.well-known/oauth-authorization-server http_status: 200 tool_catalog_source: Tool names and descriptions are taken from the published capability pages (docs.tryprofound.com/mcp/capabilities/*), not from a live tools/list — the live manifest requires authentication. Input schemas therefore require authenticated introspection and are not recorded here.