kaiten
Production-ready MCP server for Kaiten API integration with Claude Desktop
claude mcp add --transport stdio vsaranyuk-kaiten-mcp-server node /полный/путь/к/MCP Kaiten/dist/index.js \ --env KAITEN_API_URL="https://your-domain.kaiten.ru/api/latest" \ --env KAITEN_API_TOKEN="your_api_token_here" \ --env KAITEN_DEFAULT_SPACE_ID="12345"
How to use
Kaiten MCP Server integrates Kaiten API with Claude Desktop, allowing you to manage Kaiten cards, comments, and spaces directly from Claude. The server exposes a suite of tools to read, create, update, delete, and search cards; manage comments; and navigate Kaiten spaces and boards. It also supports advanced options like verbosity control, response format selection, auto-truncation for long contexts, and a production-ready feature set (validation, retries, rate limiting, caching, and detailed logging). You can perform typical workflows such as listing spaces, fetching or editing cards, adding comments, and performing filtered searches, all while controlling how much data is returned and how it’s formatted. The tools are designed to work with Claude Desktop’s config by pointing to the built MCP server executable path and, optionally, providing environment variables for Kaiten authentication and default space context.
To use, install dependencies and configure the .env or Claude config as shown in the README. Then restart Claude Desktop and invoke the Kaiten tools from Claude’s prompt. Available tools cover cards, comments, spaces/boards, users, and caching/diagnostics, with verbosity and format options to tailor the output for your workflow. For example, you can request a minimal list of cards on a board, fetch a detailed representation of a specific card, or retrieve a board in JSON for programmatic processing.
How to install
Prerequisites:
- Node.js installed (LTS version)
- npm or pnpm available
- Access to Kaiten API with a valid URL and API token
Step-by-step:
- Clone or download the Kaiten MCP server project.
- Install dependencies: npm install
- Create a proper .env file or use Claude Desktop configuration. Example .env content:
KAITEN_API_URL=https://your-domain.kaiten.ru/api/latest
KAITEN_API_TOKEN=your_api_token_here
KAITEN_DEFAULT_SPACE_ID=12345
Optional performance settings (adjust as needed)
KAITEN_MAX_CONCURRENT_REQUESTS=5 KAITEN_CACHE_TTL_SECONDS=300 KAITEN_REQUEST_TIMEOUT_MS=10000 - Build the project if a build step exists (e.g., npm run build).
- Start the MCP server locally to verify it’s listening, then configure Claude Desktop to point to the dist/index.js path as shown in the configuration example.
- In Claude Desktop, place the Kaiten MCP server configuration under mcpServers.kaiten with the path to dist/index.js and an appropriate cwd, or use the alternative env-based config if preferred.
Additional notes
Tips and notes:
- Ensure KAITEN_API_URL and KAITEN_API_TOKEN are kept secure; avoid committing .env with real tokens.
- If you use the Claude Desktop JSON config method, you can omit .env and provide env variables directly in the config as shown in the README.
- The server supports verbosity controls (minimal/normal/detailed) to optimize token usage and response size depending on the task.
- Default Space behavior means most card operations will target KAITEN_DEFAULT_SPACE_ID unless space_id is explicitly provided.
- If you encounter rate limiting or transient errors, enable exponential backoff retry settings and review the max concurrent requests.
- For debugging, use kaiten_get_status and kaiten_set_log_level to inspect cache, queue, and logs in runtime.
Related MCP Servers
zen
Selfhosted notes app. Single golang binary, notes stored as markdown within SQLite, full-text search, very low resource usage
MCP -Deepseek_R1
A Model Context Protocol (MCP) server implementation connecting Claude Desktop with DeepSeek's language models (R1/V3)
mcp-fhir
A Model Context Protocol implementation for FHIR
mcp
Inkdrop Model Context Protocol Server
mcp-appium-gestures
This is a Model Context Protocol (MCP) server providing resources and tools for Appium mobile gestures using Actions API..
dubco -npm
The (Unofficial) dubco-mcp-server enables AI assistants to manage Dub.co short links via the Model Context Protocol. It provides three MCP tools: create_link for generating new short URLs, update_link for modifying existing links, and delete_link for removing short links.