Easy English Tool 英语学习速通工具

Zero-dependency embedded MCP server for English learning — 22 tools covering vocab, articles, quizzes, copy practice, TTS, and pronunciation eval, all served directly from the Express process with no separate MCP instance. Dual transport (SSE + Streamable HTTP). Production-ready。

Easy English Tool (英语进阶营)

This skill gives you access to the user's personal English learning data and actions. The backend is an Express server serving both REST APIs and MCP endpoints on the same port — no separate MCP process needed.

Quick check: is MCP configured?

Before using this skill, check your available tools. If you see any tool named account_login, get_profile, get_today, vocab_add, etc. — the MCP server is configured. Jump to Path A. If none of these tools exist, use Path B (REST fallback).

---

Path A: MCP is configured (preferred)

The MCP server is named easy-english-tool and exposes 22 tools on https://english.geeyo.com. The MCP endpoints are served in-process by the Express backend — no separate MCP process needed.

Authentication flow

1. The user sets a username/password on the web "My" page (POST /api/users/bind).
2. Call account_login with those credentials. The MCP session caches userId, and all subsequent tools automatically use it.
3. If the user has no account yet, guide them to the web page first, then come back.

Tool reference

Account

Tool	Params	Description
account_login	username, password	Login to cache userId for the session
account_whoami	(none)	Returns cached {userId, username} or {userId: null}

Profile & Learning

Tool	Params	Description
get_profile	(none)	Learning profile: level, streak, total points, vocab count, accuracy, nickname
get_today	(none)	Today's recommended article + pending review vocab count. Returns needLevel:true if level not set.
get_history	(none)	Study history (check-ins, quizzes, learning tracks)
level_set	level	Set English level, e.g. "七年级上", "高一必修一"
levels_list	(none)	List all 12 available levels in order
diagnose	(none)	Diagnostic entry: ensures user record exists

Vocabulary

Tool	Params	Description
vocab_add	word, chinese, context?	Add a word to personal vocab list
vocab_list	(none)	List all saved vocabulary
vocab_review	id	Mark a vocab word as reviewed

Articles

Tool	Params	Description
articles_create	title, content	Create a custom study article
articles_list	(none)	List user's custom articles
articles_get	id	Get article details by id
articles_delete	id	Delete an article by id

Study Actions

Tool	Params	Description
quiz_submit	correctCount, totalCount, points?, articleId?	Submit quiz results, auto-updates points and streak
copy_load	articleId, source?	Load copy-writing record (count + draft)
copy_submit	articleId, source?	Submit a copy, count+1, auto-adds points (max 5/day)
copy_stash	articleId, draft, source?	Save a copy draft without incrementing count
evaluate	text, articleId?	Pronunciation evaluation (stub; ISE coming in next phase)

Media

Tool	Params	Description
tts	text, speed? (slow/normal/fast)	Text-to-speech, returns MP3 URL
checkin_poster	(none)	Generate check-in poster QR code (PNG)

---

Path B: MCP is NOT configured (REST fallback)

If the MCP server is not registered in your agent config, call the REST API directly at https://english.geeyo.com.

Configuring MCP (recommended)

To add the MCP server permanently, add this to your agent's MCP config:

json

{
  "mcpServers": {
    "easy-english-tool": {
      "type": "streamableHttp",
      "url": "https://english.geeyo.com/api/mcp/http"
    }
  }
}

Or for SSE-based clients:

json

{
  "mcpServers": {
    "easy-english-tool": {
      "type": "sse",
      "url": "https://english.geeyo.com/api/mcp/sse"
    }
  }
}

After adding, restart the agent and return to Path A.

REST API fallback mapping

All endpoints share the same base: https://english.geeyo.com. Anonymous auth: include header x-user-id: <userId>. Get a userId by calling POST /api/users/ensure.

MCP Tool	REST Equivalent
account_login	POST /api/users/login {username, password}
get_profile	GET /api/profile
get_today	GET /api/today
vocab_add	POST /api/vocab {word, chinese, context?}
vocab_list	GET /api/vocab
vocab_review	POST /api/vocab/review {id}
articles_create	POST /api/user-articles {title, content}
articles_list	GET /api/user-articles
articles_get	GET /api/user-articles/:id
articles_delete	DELETE /api/user-articles/:id
tts	POST /api/tts {text, speed?}
checkin_poster	GET /api/checkin/poster → PNG binary
get_history	GET /api/history
quiz_submit	POST /api/quiz {correctCount, totalCount, points?, articleId?}
level_set	POST /api/level {level}
levels_list	GET /api/levels
diagnose	POST /api/diagnose {}
copy_load	POST /api/copy {articleId, source?, action:"load"}
copy_submit	POST /api/copy {articleId, source?, action:"submit"}
copy_stash	POST /api/copy {articleId, draft, source?, action:"stash"}
evaluate	POST /api/evaluate {text, articleId?}

Server lifecycle

• Production: https://english.geeyo.com
• Local dev: cd C:\projects\Node.js\H5\server && npm run dev (listens on :3001)
• Health check: GET https://english.geeyo.com/api/health → {"ok":true}
• Zero extra deps: MCP runs in-process with Express, no @modelcontextprotocol/sdk needed.

Common workflows

First-time setup for a user:

1. Call POST /api/users/ensure with x-user-id: <uuid> → get userId
2. Call POST /api/diagnose → ensure user record
3. Call GET /api/levels → pick a level
4. Call POST /api/level {level} → set level
5. Optionally call POST /api/users/bind {username, password} → bind account for cross-device

Daily study session:

1. GET /api/today → get today's article
2. GET /api/vocab → review vocabulary
3. POST /api/tts → listen to article
4. POST /api/copy {action:"submit"} or {action:"load"} → practice writing
5. POST /api/quiz → submit quiz
6. GET /api/checkin/poster → generate sharing poster
7. GET /api/profile → check updated stats