Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Yuque MCP Server search my knowledge base for AI meeting notes"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
该项目使用AI进行重构测试 测试提交
MCP server for Yuque (语雀) — expose your knowledge base to AI assistants through the Model Context Protocol.
Visit Yuque Developer Settings to create a personal access token.
Use the built-in CLI to auto-configure your MCP client in one command:
npx yuque-mcp install --token=YOUR_TOKEN --client=cursorSupported clients: claude-desktop, vscode, cursor, windsurf, cline, trae
Or use the interactive setup wizard:
npx yuque-mcp setupThe CLI will automatically find the correct config file for your OS, merge with any existing configuration (without overwriting other servers), and print a success message.
Choose your preferred client below:
claude mcp add yuque-mcp -- npx -y yuque-mcp --token=YOUR_TOKENOr using environment variables:
export YUQUE_PERSONAL_TOKEN=YOUR_TOKEN
claude mcp add yuque-mcp -- npx -y yuque-mcpAdd to your claude_desktop_config.json:
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"yuque": {
"command": "npx",
"args": ["-y", "yuque-mcp"],
"env": {
"YUQUE_PERSONAL_TOKEN": "YOUR_TOKEN"
}
}
}
}Add to .vscode/mcp.json in your workspace:
{
"servers": {
"yuque": {
"command": "npx",
"args": ["-y", "yuque-mcp"],
"env": {
"YUQUE_PERSONAL_TOKEN": "YOUR_TOKEN"
}
}
}
}Then enable Agent mode in GitHub Copilot Chat.
Add to your Cursor MCP configuration (~/.cursor/mcp.json):
{
"mcpServers": {
"yuque": {
"command": "npx",
"args": ["-y", "yuque-mcp"],
"env": {
"YUQUE_PERSONAL_TOKEN": "YOUR_TOKEN"
}
}
}
}Add to your Windsurf MCP configuration (~/.windsurf/mcp.json):
{
"mcpServers": {
"yuque": {
"command": "npx",
"args": ["-y", "yuque-mcp"],
"env": {
"YUQUE_PERSONAL_TOKEN": "YOUR_TOKEN"
}
}
}
}Add to your Cline MCP settings (~/Library/Application Support/Code/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json):
{
"mcpServers": {
"yuque": {
"command": "npx",
"args": ["-y", "yuque-mcp"],
"env": {
"YUQUE_PERSONAL_TOKEN": "YOUR_TOKEN"
}
}
}
}In Trae, open Settings and navigate to the MCP section, then add a new stdio-type MCP Server with the following configuration:
Command: npx
Args: -y yuque-mcp
Env: YUQUE_PERSONAL_TOKEN=YOUR_TOKEN
See Trae MCP documentation for detailed instructions.
More clients: Any MCP-compatible client that supports stdio transport can use yuque-mcp. The general pattern is: command =
npx, args =["-y", "yuque-mcp"], env =YUQUE_PERSONAL_TOKEN.
Ask your AI assistant to search your Yuque docs, create documents, or manage books.
Related MCP server: Yuque MCP Server
The server supports multiple ways to provide your Yuque API token:
Method | Environment Variable / Flag | Description |
Personal Token (recommended) |
| For accessing your personal Yuque account |
Group Token |
| For accessing a Yuque group |
Legacy Token |
| Backward-compatible, works the same |
CLI Argument |
| Pass directly as a command-line argument |
Priority order: YUQUE_PERSONAL_TOKEN > YUQUE_GROUP_TOKEN > YUQUE_TOKEN > --token
Category | Tools |
User |
|
Search |
|
Books |
|
Docs |
|
TOC |
|
Versions |
|
Groups |
|
Stats |
|
Utility |
|
Error | Solution |
| Set one of the environment variables ( |
| Token is invalid or expired — regenerate at Yuque Settings |
| Too many requests — wait a moment and retry |
Tool not found | Update to the latest version: |
| Install Node.js (v18 or later) |
git clone https://github.com/yuque/yuque-mcp-server.git
cd yuque-mcp-server
npm install
npm test # run tests
npm run build # compile TypeScript
npm run dev # dev mode with hot reloadCreate a new document in a repo/book
Inputs: title (required), repo_id (required), body, slug, +2 more.
Create a new repo/book for a user or group
Inputs: name (required), slug (required), type (required), login (required), +3 more.
Delete a document
Inputs: doc_id (required), repo_id (required).
Delete a repo/book
Inputs: id_or_namespace (required).
Get a specific document with full content
Inputs: doc_id (required), repo_id (required).
Get a specific version of a document
Inputs: version_id (required).
Get a specific repo/book by ID or namespace (group_login/book_slug)
Inputs: id_or_namespace (required).
Get the table of contents (TOC) for a repo/book
Inputs: repo_id (required).
Get current authenticated user information
Inputs: no named parameters.
Get book/repo statistics for a group/team
Inputs: login (required).
Get document statistics for a group/team
Inputs: login (required).
Get member statistics for a group/team
Inputs: login (required).
Get overall statistics for a group/team
Inputs: login (required).
Test API connectivity with Yuque
Inputs: no named parameters.
List all documents in a repo/book
Inputs: repo_id (required).
List all versions of a document
Inputs: doc_id (required).
List all members of a group/team
Inputs: login (required).
List all groups/teams that the user belongs to
Inputs: user_id (required).
List all repos/books for a user or group
Inputs: type (required), login (required).
Remove a member from a group/team
Inputs: login (required), user_id (required).
Search for documents, repos, or users in Yuque
Inputs: query (required), type.
Update an existing document
Inputs: doc_id (required), repo_id (required), body, slug, +2 more.
Update a group member role
Inputs: role (required), login (required), user_id (required).
Update a repo/book
Inputs: id_or_namespace (required), name, slug, public, +1 more.
Update the table of contents (TOC) for a repo/book
Inputs: repo_id (required), toc_data (required).
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 25 tools
Most tools are distinct (get vs list vs create vs delete), but the four group stats tools (group_stats, group_member_stats, group_book_stats, group_doc_stats) have overlapping names and could be confused for one another. All other tools have clearly separated purposes.
The dominant pattern is yuque_verb_noun (e.g., get_repo, create_doc, list_docs). The four group stats tools break this by using group_..._stats without a verb, and yuque_hello is an outlier. Overall, naming is mostly consistent with minor deviations.
With 25 tools, this is at the upper boundary of a heavy toolset. The domain covers repos, docs, groups, members, versions, TOC, stats, and search, but several stats tools could be consolidated. It is not bloated, but it is slightly over the typical well-scoped range.
Repos and docs have full CRUD coverage, plus TOC, versions, and search. However, there is no group creation/deletion, no way to add a member (only remove/update), and no features like comments or attachments. These are notable gaps for a collaborative knowledge base platform.
Search and reason over your Obsidian-style Markdown vault, right from ChatGPT.
Personal context for every AI: search, read, and write back to your private Markdown library.
Read and write KukGit repositories, files, issues and pull requests from an AI assistant.
Gives AI agents direct access to your Sanity projects and guidance for working with them
