generated: '2026-09-25' method: probed source: - https://framethrower.ai/api/mcp - https://framethrower.ai/mcp - https://github.com/framethrower-ai/framethrower-mcp - https://registry.modelcontextprotocol.io/v0/servers?search=framethrower status: live server: name: FrameThrower version: 1.0.0 protocolVersion: '2025-06-18' registry_name: ai.framethrower/framethrower transport: streamable-http capabilities: tools: listChanged: true auth: type: oauth2 note: OAuth 2.1 per the provider README; authorization server metadata at https://framethrower.ai/.well-known/oauth-authorization-server (authorization_code + PKCE S256, dynamic client registration at /api/auth/mcp/register). initialize and tools/list answered anonymously (200) on 2026-09-25; tool calls spend account credits. tools: - name: search_frames inputs: - query - limit required: - query description: Search the FrameThrower cinematography reference library by concept, mood, color, composition, or scene. Returns reference frames with film/credits metadata, a preview thumbnail, the full-resolution image URL (use this one for slide decks and documents), a precomputed DEPTH MAP url (greyscale, same dimensions as the frame — pass it straight to a depth ControlNet or any image-to-image pipeline that… - name: find_by_craft inputs: - lens - shot_size - visual_style - setting - time_of_day - director - year_min - year_max - limit description: Find frames by specific cinematography craft attributes — lens character, shot size, visual style, setting, time of day, director, or year range. All optional; combine to narrow. Use this to refine a search precisely. - name: find_similar inputs: - frameId - limit required: - frameId description: Find frames visually/semantically similar to a given frame id (from a prior result). Use this to pull more references like one the user shortlisted. - name: get_frame_details inputs: - frameId required: - frameId description: Get the full metadata + preview thumbnail + full-resolution image URL + depth map URL + link for a single frame id. The depth map is precomputed and free — use it when the user is generating images and needs the composition or camera geometry of this frame. tools_list_file: mcp/framethrower-mcp-tools-list.json deployment: mode: remote endpoint: https://framethrower.ai/api/mcp auth: oauth verified: probed note: The README offers npx -y mcp-remote https://framethrower.ai/api/mcp as a bridge for stdio-only clients; mcp-remote is a third-party bridge, not a FrameThrower package. descriptions_source: live tools/list (authorship probe or saved capture); filled by backfill-tools-from-authorship.py --descriptions, 2026-09-28