generated: '2026-08-04' method: searched status: published source: https://help.clozd.com/hc/en-us/articles/49656607624987-Connecting-to-Clozd-via-MCP also_documented_at: https://www.clozd.com/product/platform/clozd-mcp server: name: clozd transport: http url: https://mcp.clozd.com/mcp protocol: Model Context Protocol (streamable HTTP) auth: type: oauth2 flow: authorization_code pkce: S256 authorization_server: https://oauth.clozd.com dynamic_client_registration: true scopes: - openid - profile - email - offline_access - api identity_brokering: - Okta - Google - Microsoft / Entra - SAML - OneLogin - email + password notes: >- No API keys or tokens are placed in client configuration. The AI client opens a browser window and the user signs in through their organization's identity provider. Role-based access control and program scoping are enforced by the Clozd platform on the authenticated user. live_probe: fetched: '2026-08-04' method: 'POST /mcp {"jsonrpc":"2.0","id":1,"method":"tools/list"}' http_status: 401 body: '{"success":false,"message":"MCP authentication failed.","errorCode":"MCP001"}' www_authenticate: >- Bearer realm="Clozd MCP", resource_metadata="https://mcp.clozd.com/.well-known/oauth-protected-resource/mcp" conclusion: >- tools/list is auth-gated, so live inputSchema introspection is not possible anonymously. The tool list below is transcribed verbatim from Clozd's published help-center table and product page — names and descriptions only. Input schemas require an authenticated introspection pass. supported_clients: - Claude (web, Desktop, Cowork) — custom connector - ChatGPT — Apps directory - Cursor - Microsoft Copilot Studio - Windsurf - Antigravity - Gemini CLI - any MCP-compatible client via URL or JSON config tool_count: 19 tools: - name: get_programs category: programs description: >- Lists the Clozd win-loss programs available to the authenticated user. Call this first to retrieve program IDs needed by all other tools. - name: get_deals category: deals description: >- Returns a paginated list of deals with fields like outcome, amount, industry, closed date, sales rep, region, and segment. Use to browse or search deal records. - name: get_responses category: feedback description: >- Returns a paginated list of win-loss interview and survey responses, including deal link, respondent role, channel, and publish date. - name: get_response_summaries category: feedback description: >- Returns AI-generated summaries of published win-loss feedback. Faster than loading full transcripts when you need a quick read on a response. - name: get_transcripts category: feedback description: >- Returns full-text transcripts of win-loss interviews or surveys. Best used when the complete conversation content is needed; prefer summaries or quotes for most use cases. - name: get_decision_drivers category: drivers description: >- Lists individual decision drivers ranked by mention count with sentiment breakdowns. Use to answer "What are the most common reasons we win or lose?" - name: get_decision_driver_categories category: drivers description: >- Lists high-level driver categories (e.g., Product, Pricing, Support) with aggregated sentiment. Use for broad thematic questions like "Which category of feedback is most negative?" - name: get_decision_driver_category_counts category: drivers description: >- Returns a frequency count of driver mentions by category. Use when you want a ranked summary of which categories appear most. - name: get_driver_quotes category: drivers description: >- Returns verbatim buyer quotes linked to a specific decision driver or theme, including respondent context. Use for direct evidence (e.g., "Show me quotes about pricing.") - name: get_win_rates category: analytics description: >- Calculates overall win-loss rates with optional breakdown by dimension (industry, product, sales rep, deal size, etc.). - name: get_competitors category: competitive description: >- Lists competitors encountered across deals with encounter counts and win-loss tallies per competitor. - name: get_competitor_sentiment_drivers category: competitive description: >- Returns the decision drivers associated with a specific competitor filtered by positive or negative sentiment. Use to answer "Why do we beat Competitor X?" or "Why do we lose to them?" - name: get_awe_deals category: deals description: >- Returns deals flagged as At-risk, Win-back, or Expansion opportunities. Use to surface actionable accounts for CS, sales, or revenue teams. - name: get_tags category: metadata description: >- Lists user-defined tags applied to deals with usage counts. Optionally returns the deal IDs associated with each tag. - name: get_gong_drivers category: gong description: >- Lists AI-detected decision drivers surfaced from Gong call recordings, with sentiment distribution. Distinct from interview-based drivers. - name: get_gong_driver_categories category: gong description: >- Lists high-level category groupings of Gong-identified drivers with sentiment. Use for broad questions about themes emerging in sales calls. - name: get_gong_driver_counts category: gong description: >- Ranks Gong-identified drivers by frequency of mention. Use to find the most common themes across Gong call data. - name: get_gong_driver_category_counts category: gong description: >- Returns a frequency count of Gong driver mentions by category. Use for a ranked category-level view of Gong call themes. - name: get_gong_driver_overlap category: gong description: >- Compares Gong-identified drivers against Clozd interview drivers to show alignment and gaps between the two data sources. tools_listed_on_product_page_only: - name: get_decision_driver_counts note: >- Listed on https://www.clozd.com/product/platform/clozd-mcp but absent from the help-center tool table (which lists get_decision_driver_category_counts instead). Recorded as a discrepancy between the two published sources, not asserted as a live tool. roadmap: source: https://help.clozd.com/hc/en-us/articles/49656607624987-Connecting-to-Clozd-via-MCP items: - Native Clozd MCP connector submitted to the Claude.ai connector marketplace - Native Clozd MCP listing submitted to the OpenAI plugin directory support: email: support@clozd.com help_center: https://help.clozd.com/hc/en-us x-evidence: fetched: '2026-08-04' urls: - url: https://help.clozd.com/api/v2/help_center/en-us/articles/49656607624987.json http_status: 200 - url: https://www.clozd.com/product/platform/clozd-mcp http_status: 200 - url: https://mcp.clozd.com/mcp http_status: 401 - url: https://mcp.clozd.com/.well-known/oauth-protected-resource/mcp http_status: 200