Skip to main content
Glama

anki-librarian-mcp

See it working · Overview, 1 page (PDF)

Public version of the flashcard workflow I use in my own studies: Claude turns study material into Anki cards and manages the collection, without ever writing anything you have not approved.

Built with AI assistance Fictitious demo data

The two examples, side by side: material, debrief, plan

What it shows

  • From notes to cards in one request. Claude sums up what the material teaches, plans the cards, writes them and tags them, shown step by step for dense econometrics notes and for a business case.

  • No duplicates. Before planning, it searches your whole collection and shows the card you already have.

  • Nothing is written without you. Every change is a plan you approve first; deleting and rescheduling stay off unless you turn them on.

Related MCP server: enterprise-agent-lab

Run it

pip install -r requirements.txt
python demo.py

Runs without an API key, replaying recorded answers. Set ANTHROPIC_API_KEY and add --live to call Claude live. It uses a practice collection; your Anki is not touched.

How I use it

Most of my cards come from a weekly radar of news and data by topic, prepared by a scheduled task on Thursday night; I pick from its menu and approve the cards. The rest come from quick questions, the end of a project and dense study material. The radar does not run by itself: you ask for it or schedule it (how).

More

  • How it works: how I use it, each step, the safety rules, installing it for Claude Desktop or Claude Code, and how to start on your own collection with a test deck.

  • Runbook: commands, updating, and what to do when something breaks.

  • The other side of the flashcards, how much review time each deck needs in the week, is in personal-ai-agents (study blocks).

Limits

  • Pictures only, no audio or video.

  • "Similar" means shared words and phrases; a card on the same idea in completely different words can be missed.

  • Tested against a practice collection that behaves like Anki; try it on a test deck of your own first.

How AI was used

I use Claude (Claude Code and the API) for coding, review and documentation. The questions, modelling choices, validation and conclusions are mine.

License

MIT, Guilherme Cabral.

Available Tools

13 tools
apply_planA
DestructiveIdempotent

Write an approved plan to Anki and read it back to confirm. If the plan is still pending, the server asks the user to approve it (the model cannot approve on the user's behalf).

Inputs: plan_id (required).

find_notesA
Read-only

Search notes with Anki search syntax (e.g. 'deck:Economics tag:Economics::Macro inflation'). Returns note ids, model, decks, tags and field excerpts.

Inputs: query (required), limit.

get_notesC
Read-only

Full fields and tags of up to 100 notes.

Inputs: note_ids (required). Returns: result

list_decksA
Read-only

List all decks and which of them this server may write to.

Inputs: no named parameters.

list_plansC
Read-only

Plans and their status (pending, approved, applied, rejected, expired, invalid, failed).

Inputs: status. Returns: result

propose_add_cloze_notesB
Read-only

Dry run for adding cloze notes (one or many); cloze syntax is validated for the whole batch.

Inputs: deck (required), notes (required).

propose_add_notesA
Read-only

Dry run for adding basic (front/back) notes, one or many. Searches the collection first: exact duplicates are skipped, 'same front, different back' is left for the user, similar fronts are flagged, tags are checked against the taxonomy. Nothing is written: call apply_plan after the user sees it.

Inputs: deck (required), notes (required).

propose_create_deckC
Read-only

Dry run for creating a deck (only inside the writable decks).

Inputs: name (required).

propose_delete_notesA
Read-only

Dry run for deleting notes (and their review history). Off unless the user enabled it in the config; at most 50 ids, no delete-by-query.

Inputs: note_ids (required).

propose_flag_cardsA
Read-only

Dry run for flagging cards with an agent color (default: 4 blue, or 0 to clear one the agent set). Cards carrying a flag the user set are skipped, never overwritten.

Inputs: color (required), card_ids (required).

propose_set_due_dateA
Read-only

Dry run for rescheduling cards (Anki's set due date, e.g. days='0' or '1-3'). Off unless enabled.

Inputs: days (required), card_ids (required).

propose_update_noteA
Read-only

Dry run for editing one note's fields and/or replacing its tags. Shows a before/after diff. If the note changes in Anki before apply_plan, the plan is refused (no stale writes).

Inputs: note_id (required), tags, fields.

reject_planC

Discard a plan the user does not want.

Inputs: plan_id (required), reason.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A3.6/5.0

    Scored across 13 tools

    Disambiguation5/5

    Each tool targets a distinct action and resource: read tools (list_decks, find_notes, get_notes), propose tools for different mutations (add notes, add cloze, update, create deck, delete, set due date, flag), and plan lifecycle tools (list_plans, apply_plan, reject_plan). find_notes (search excerpts) vs get_notes (full fields) are clearly separated, and the propose_* tools differ by object and operation.

    Naming Consistency4/5

    Strong, predictable verb_noun pattern throughout, with a consistent propose_* prefix for dry runs and matching apply_plan/reject_plan lifecycle verbs. Minor deviations like singular vs plural (propose_update_note vs propose_add_notes) keep it just short of perfect.

    Tool Count5/5

    13 tools is well-scoped for the Anki domain. The propose/apply split is intentional and each tool earns its place by covering a specific mutation or query without redundancy.

    Completeness4/5

    Covers a full lifecycle: deck/note discovery, note and cloze creation, updates, deletion, scheduling, flagging, and plan management. Minor gaps like moving notes between decks or deleting decks are absent but not blocking for core workflows.

    Maintenance

    ActivityMaintained
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      B
      quality
      B
      maintenance
      Enables LLM agents to query databases with read-only access, while requiring human approval for writes through a token-based confirmation system.
      6
      GPL 3.0
    • F
      license
      Not graded
      quality
      C
      maintenance
      Enables controlled AI-agent access to enterprise-shaped tools with a deny-by-default gated write path, human approval, dry-run execution, and append-only audit logging.
      1
      -
    • A
      license
      A
      quality
      C
      maintenance
      Enables AI agents and assistants to work with a Databricks workspace — SQL, clusters and warehouses, notebooks, Jobs, Lakeflow pipelines, Unity Catalog, Volumes, dashboards, Genie, model serving, Vector Search, Lakebase and Apps — under a safety-first model with read/write/destructive/security classification, confirmation steps for dangerous changes, production resource protection, and a fully read-only mode. Every call is checked against the official Databricks SDK, returns typed schemas and redacted secrets, and unsupported features report an explicit error rather than faking results.
      45
      Apache 2.0
    • A
      license
      A
      quality
      B
      maintenance
      Enables LLMs to safely explore, query, diagnose, and modify relational databases through a unified set of schema introspection, SQL execution, and administration tools. Supports multiple database backends with AST-level read-only guards, automatic LIMIT injection, and explicit opt-in write permissions for DML and DDL.
      6
      1
      Apache 2.0