# Vendor facets — Toolhouse. Toolhouse has pivoted to "AI workers": it is an MCP CLIENT that # connects workers to external MCP servers and its own built-in integrations, and exposes each # worker as an API at agents.toolhouse.ai. It does not host MCP servers for API providers, so # nearly everything it ships earns a provider nothing. vendor: toolhouse name: Toolhouse website: https://toolhouse.ai areas: - mcp-hosting registry_keys: [] rubric_schema_version: 0.22.0 generated: '2026-09-25' features_refreshed: '2026-09-25' basis: capability summary: >- Toolhouse no longer fits mcp-hosting from a provider's point of view: its current product is AI workers that consume integrations and connect out to MCP servers other people host (its "official partner" Tavily is wired in via Tavily's own mcp.tavily.com). A provider gains Kin Score only from its own MCP server, which Toolhouse neither builds nor hosts; a Toolhouse integration for your API earns nothing. The one qualified row is a provider-run MCP server that Toolhouse can consume — the credit is the provider's, not Toolhouse's. features: - id: mcp-client name: Connect workers to external MCP servers description: >- Workers attach external MCP servers over SSE or Streamable HTTP with OAuth 2.0/2.1 or path/query auth; no header auth. source: https://docs.toolhouse.ai/toolhouse/toolhouse/mcp-servers.md tier: all - id: integrations name: Built-in and connection-required integrations description: Toolhouse-built integrations and third-party connections the worker picks automatically. source: https://docs.toolhouse.ai/toolhouse/toolhouse/integrations.md tier: all - id: workers-api name: Workers API description: Each worker is callable at https://agents.toolhouse.ai/ with an API key. source: https://docs.toolhouse.ai/toolhouse/developers/workers-api.md tier: all - id: agent-files name: Skills and knowledge files description: Users upload SKILLS.md, templates and guidelines into a worker's instructions. source: https://docs.toolhouse.ai/toolhouse/capabilites/skills-and-knowledge.md tier: all - id: partner-listing name: Official partner integration pages description: Partner pages (e.g. Tavily) show how to add the partner's own hosted MCP URL to a worker. source: https://docs.toolhouse.ai/toolhouse/integrations/tavily-official-partner.md tier: unknown maps: [] earns_nothing: - feature: integrations check: mcp_server why: >- A Toolhouse-built integration for a provider's API runs inside Toolhouse workers; it is not an MCP server the provider publishes. - feature: agent-files check: agent_skills why: >- Skill files are uploaded by the Toolhouse user into a private worker; they are not provider-authored, published skills. - feature: workers-api check: agent_card why: >- Worker endpoints on agents.toolhouse.ai carry no A2A agent card, and they belong to the Toolhouse user, not the provider. - feature: mcp-client check: protected_resource_metadata why: Toolhouse is an OAuth client of MCP servers; it serves no protected-resource metadata for anyone. out_of_reach: checks: - contract_present - spec_presence - llms_txt_published - well_known_catalog - dynamic_client_registration - agent_skill_present note: Toolhouse consumes APIs and MCP servers; nothing it ships lands on a provider's surface. surface: agent_readiness: reachable: 0 total: 139 hard_rule: >- A model, not a score. Adopting this vendor changes a provider's Kin Score only when the provider publishes the resulting artifacts on its own surface; nothing here writes a score, and no sponsorship or partnership can. method: searched source: - https://docs.toolhouse.ai/toolhouse/capabilites/skills-and-knowledge.md - https://docs.toolhouse.ai/toolhouse/developers/workers-api.md - https://docs.toolhouse.ai/toolhouse/integrations/tavily-official-partner.md - https://docs.toolhouse.ai/toolhouse/toolhouse/integrations.md - https://docs.toolhouse.ai/toolhouse/toolhouse/mcp-servers.md measured: cohort: method: vendors-catalog.json detections (CNAME / header / URL shape / markup), never a name match detected: 0 in_baseline: 0 control: basis: providers earning contract_present + documentation_present + api_reference_present, minus the cohort n: 5216 metric: >- cohort_pct / control_pct = mean share of the check's points earned (derived and platform credit weighted), x100 measured_on: '2026-09-25' status: 'not measurable: 0 detected customers clear the baseline (need 20)' simulation: simulated_on: '2026-09-25' rubric: 0.23.0 population: providers publishing a contract (contract_present earned), replayable exactly providers: 8977 providers_unreplayable: 987 providers_moved: 0 conditional_rows: excluded (they depend on what the API already does) composite_lift: median: 0.0 p75: 0.0 p90: 0.0 max: 0.0 mean_among_movers: 0.0 agent_readiness_lift: median: 0.0 p75: 0.0 p90: 0.0 max: 0.0 mean_among_movers: 0.0 facet_lift_median_among_movers: {} composite_band_moves: {} agent_readiness_band_moves: {} method: >- each provider's own kin/checks file, the vendor's maps at their stated credit, the scorer's composite formula; from -> to, nothing written