mcp-server-maimemo
Exposes the Maimemo (墨墨背单词) Open API as MCP tools, enabling AI assistants to look up words, manage mnemonics and study plans, and maintain cloud notepads.
README
mcp-server-maimemo
An MCP (Model Context Protocol) server for the 墨墨背单词 (Maimemo) Open API. It exposes every documented API capability as a typed MCP tool, so AI assistants can look up words, manage mnemonics / interpretations / example sentences, maintain cloud notepads, and read or modify study plans.
Highlights
- Complete API coverage — 24 tools spanning all six documented resource groups.
- Streamable HTTP transport — the current MCP spec, at
POST /mcp. The deprecated HTTP+SSE transport is still served atGET /ssefor older clients. - Per-request tokens — clients may send
Authorization: Bearer <maimemo_token>, so one deployment can safely serve multiple users. Each session gets an isolated API client. - Built-in rate limiting — a multi-window sliding log shapes traffic to Maimemo's quotas (10s/20, 60s/40, 5h/2000 by default) instead of letting the model hit random 429s.
- Retries, timeouts, and caching — exponential backoff with jitter on transient failures,
per-request timeouts, and a bounded
spelling → voc_idcache that removes redundant lookups. - Schema-validated inputs — tools are declared once as zod schemas and projected into JSON Schema, so advertised contracts and runtime validation cannot drift apart.
- Actionable errors — failures come back as typed, human-readable
isErrorresults (auth,rate_limit,not_found,invalid_input, …) so the model can self-correct.
Tools
| Group | Tools |
|---|---|
| Vocabulary 单词 | get_vocabulary, query_vocabularies |
| Notes 助记 | list_notes, create_note, update_note, delete_note |
| Interpretations 释义 | list_interpretations, create_interpretation, update_interpretation, delete_interpretation |
| Phrases 例句 | list_phrases, create_phrase, update_phrase, delete_phrase |
| Notepads 云词本 | list_notepads, get_notepad, create_notepad, update_notepad, delete_notepad |
| Study 学习数据 | get_study_progress, get_today_items, query_study_records, add_words, advance_study |
Word-scoped tools accept either word (a spelling, resolved automatically and cached) or a
raw voc_id. Read-only tools are annotated with readOnlyHint, and delete tools with
destructiveHint, so clients can gate them appropriately.
Getting a token
In the Maimemo app: 我的 → 更多设置 → 实验功能 → 开放 API
Quick start
git clone https://github.com/Mrquj/mcp-server-maimemo.git
cd mcp-server-maimemo
npm install
npm run build
cp .env.example .env # then fill in MAIMEMO_API_TOKEN
npm start
Verify it is up:
curl http://localhost:3000/healthz
Docker
docker build -t mcp-server-maimemo .
docker run -p 3000:3000 -e MAIMEMO_API_TOKEN=your_token mcp-server-maimemo
Configuration
| Variable | Default | Purpose |
|---|---|---|
MAIMEMO_API_TOKEN |
— | Maimemo Open API token. Required for stdio; optional for http when clients send their own. |
MAIMEMO_REQUIRE_CLIENT_TOKEN |
true if no token is set |
Force every client to send Authorization: Bearer. |
MAIMEMO_MCP_TRANSPORT |
http |
http or stdio. |
PORT / HOST |
3000 / 0.0.0.0 |
HTTP listener. |
MAIMEMO_REQUEST_TIMEOUT_MS |
20000 |
Per-request timeout. |
MAIMEMO_MAX_RETRIES |
3 |
Retry attempts for transient failures. |
MAIMEMO_RATE_LIMITS |
10s:20,60s:40,5h:2000 |
Client-side throttling windows. |
MAIMEMO_API_BASE_URL |
https://open.maimemo.com/open/ |
Override the API base. |
Invalid configuration fails fast at startup with an explicit message.
Connecting clients
Remote clients (Notion, and other hosted MCP clients)
Deploy behind HTTPS, then register the MCP server URL:
https://your-domain.example.com/mcp
If you set MAIMEMO_REQUIRE_CLIENT_TOKEN=true, configure the connection's Bearer token to
your Maimemo API token. Otherwise the server falls back to MAIMEMO_API_TOKEN.
Local clients (Claude Desktop, and similar)
{
"mcpServers": {
"maimemo": {
"command": "node",
"args": ["/absolute/path/to/mcp-server-maimemo/dist/index.js"],
"env": {
"MAIMEMO_MCP_TRANSPORT": "stdio",
"MAIMEMO_API_TOKEN": "your_token_here"
}
}
}
}
MCP Inspector
npm run inspect
Endpoints
| Method | Path | Purpose |
|---|---|---|
POST |
/mcp |
Streamable HTTP transport (initialize + requests). |
GET |
/mcp |
Server-to-client notification stream for an existing session. |
DELETE |
/mcp |
Terminate a session. |
GET |
/sse |
Deprecated HTTP+SSE transport. |
POST |
/messages |
Message sink for the deprecated SSE transport. |
GET |
/healthz |
Liveness, version, session count, rate-limit usage. |
Example calls
{ "name": "get_vocabulary", "arguments": { "word": "apple" } }
{ "name": "create_note", "arguments": { "word": "apple", "note_type": "谐音", "note": "阿派" } }
{ "name": "add_words", "arguments": { "spellings": ["apple", "banana"], "advance": true } }
{ "name": "get_study_progress", "arguments": {} }
Notepad content uses Maimemo's markup:
#DESC
Words from this week's reading
#LIST
apple
banana
cherry
Architecture
src/
config.ts Environment parsing and fail-fast validation
errors.ts Typed error hierarchy and HTTP status mapping
rate-limiter.ts Multi-window sliding-log throttle
client.ts Maimemo API client: retries, timeouts, voc_id cache
tools.ts zod schemas → JSON Schema + tool annotations
handlers.ts Validated dispatch to Maimemo endpoints
server.ts MCP server factory
http.ts Streamable HTTP + legacy SSE + health, graceful shutdown
index.ts Entry point and transport selection
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.