Model Context Protocol (MCP) server implementation for Gitee – exposes a CLI/stdio/HTTP server with a rich tool-set for managing Gitee repos, issues, PRs, releases and notifications.
https://github.com/oschina/mcp-giteeYour AI coding assistant just got direct access to Gitee's complete API. No more tab switching, no more breaking flow to create issues or review PRs in a browser.
You're deep in a coding session with Claude or Cursor, discussing implementation details, when you need to:
Each action means leaving your AI conversation, opening Gitee in a browser, clicking through the interface, then returning to explain what you found. Your flow is broken, context is lost.
This MCP server puts 27 Gitee operations directly in your AI assistant's toolkit. Your assistant can now:
Repository Management
Issue Workflow
Pull Request Operations
User & Notification Management
The assistant handles the API calls, parses responses, and presents information in context – all without you leaving the conversation.
Here's how issue management changes:
Before: "I found a bug in the authentication flow" → Open Gitee → Navigate to Issues → Click New Issue → Fill form → Submit → Return to AI → Explain what you created
After: "I found a bug in the authentication flow" → Assistant: "Let me create an issue for that. What repository and what details should I include?" → Issue created, tracked, and referenced in your ongoing conversation
Same for PRs: "This feature is ready for review" → Assistant creates the PR, sets reviewers, adds description, and gives you the PR link → Later: "Merge that PR" → Done, with merge commit details returned
Zero Install (Recommended)
{
"mcpServers": {
"gitee": {
"url": "https://api.gitee.com/mcp",
"headers": {
"Authorization": "Bearer YOUR_TOKEN"
}
}
}
}
Connect directly to the hosted service. No local installation, no version management.
NPX for Latest Features
{
"mcpServers": {
"gitee": {
"command": "npx",
"args": ["-y", "@gitee/mcp-gitee@latest"],
"env": {
"GITEE_ACCESS_TOKEN": "YOUR_TOKEN"
}
}
}
}
Always pulls the latest version. Perfect for rapid feature updates.
Enable only what you need:
--enabled-toolsets="list_repo_issues,create_issue,create_pull,merge_pull"
Or disable specific tools:
--disabled-toolsets="create_user_repo,fork_repository"
This is particularly useful in team environments where you want to restrict certain operations or when working with organization repositories where permissions vary.
Works seamlessly with:
Each has specific configuration guides included – no guesswork on setup.
If you're working with Gitee repositories – whether for compliance reasons, team preferences, or because you're building for the Chinese market – this server bridges the gap between international AI tools and China's premier code hosting platform.
The server handles Gitee's specific API patterns, authentication requirements, and response formats, so your AI assistant works as naturally with Gitee as it does with GitHub.
Your AI assistant is about to become significantly more useful for actual development work.