{"id":"ankimcp-anki-mcp-server-addon","name":"anki-mcp-server-addon","homepage":"https://ankimcp.ai","repo_url":"https://github.com/ankimcp/anki-mcp-server-addon","category":"devtools","subcategories":[],"tags":["anki","mcp","ai","spaced-repetition","local-server","python"],"what_it_does":"Provides an Anki add-on that runs a local Model Context Protocol (MCP) server inside Anki (via HTTP transport) so AI assistants can query and manage a user's Anki collection (e.g., decks, notes, cards, tags, media) and guide review sessions.","use_cases":["AI-assisted study/review workflows using an MCP client","AI-driven deck/note/card creation and management in Anki","Searching notes using Anki's query syntax","Automating common spaced-repetition operations (rate cards, reschedule, tag management)","Tunneling/inspecting the local MCP endpoint via tools like ngrok/Cloudflare (with optional custom path)"],"not_for":["Use cases requiring strong remote multi-tenant access controls (this is primarily local/tunnel exposure)","Scenarios where users cannot trust a local service that can manipulate their Anki collection","Environments needing a public, internet-facing API with robust authentication built in"],"best_when":"You want a local, MCP-compatible bridge between an AI assistant and a single user's Anki instance running on the same machine (or safely via a tunnel with restricted exposure).","avoid_when":"You need fine-grained security/audit controls for remote access or you cannot restrict who can reach the local HTTP endpoint.","alternatives":["AnkiConnect (HTTP API for Anki)","Other local desktop automation bridges that expose Anki functionality via REST","Custom MCP server implementations that call Anki APIs directly"],"af_score":56.2,"security_score":35.0,"reliability_score":37.5,"package_type":"mcp_server","discovery_source":["github"],"priority":"high","status":"evaluated","version_evaluated":null,"last_evaluated":"2026-03-30T15:29:09.288996+00:00","interface":{"has_rest_api":false,"has_graphql":false,"has_grpc":false,"has_mcp_server":true,"mcp_server_url":"http://127.0.0.1:3141/","has_sdk":false,"sdk_languages":[],"openapi_spec_url":null,"webhooks":false},"auth":{"methods":[],"oauth":false,"scopes":false,"notes":"README describes a local HTTP MCP server with host/port/path and CORS settings but does not mention authentication (API keys, OAuth, etc.). Access control is therefore likely determined by network exposure (localhost vs tunnel) and CORS."},"pricing":{"model":null,"free_tier_exists":false,"free_tier_limits":null,"paid_tiers":[],"requires_credit_card":false,"estimated_workload_costs":null,"notes":"No pricing information provided; Anki add-ons are typically installed locally."},"requirements":{"requires_signup":false,"requires_credit_card":false,"domain_verification":false,"data_residency":[],"compliance":[],"min_contract":null},"agent_readiness":{"af_score":56.2,"security_score":35.0,"reliability_score":37.5,"mcp_server_quality":78.0,"documentation_accuracy":80.0,"error_message_quality":0.0,"error_message_notes":null,"auth_complexity":25.0,"rate_limit_clarity":20.0,"tls_enforcement":35.0,"auth_strength":15.0,"scope_granularity":20.0,"dependency_hygiene":55.0,"secret_handling":60.0,"security_notes":"Likely primarily local access (127.0.0.1) but can be tunnel-exposed. README does not describe authentication/authorization, so security depends on network restrictions and CORS. Custom http_path and CORS origins can reduce accidental exposure, but there is no mention of protecting against direct requests to the MCP endpoint or enforcing per-client authorization.","uptime_documented":20.0,"version_stability":45.0,"breaking_changes_history":35.0,"error_recovery":50.0,"idempotency_support":"false","idempotency_notes":"The README documents tool capabilities and batch behavior, but does not state idempotency guarantees for operations like add/update/delete.","pagination_style":"none","retry_guidance_documented":false,"known_agent_gotchas":["Operations can mutate or delete collection data; agents should be cautious with destructive tools (delete_notes, delete_media_file, forget/suspend/sync).","Disabled tools are removed from the MCP schema; typos in tool/action names produce console warnings (agents should validate tool names).","First run downloads pydantic_core (~2MB); agents/workflows that rely on cold-start should allow for this delay."]}}