kinaxis-mcp-server

# kinaxis-mcp-server [](LICENSE) The **first open-source** [Model Context Protocol (MCP)](https://modelcontextprotocol.io) server for [Kinaxis Maestro](https://www.kinaxis.com) (formerly RapidResponse). > **Note:** Genpact offers a proprietary Kinaxis MCP integration, but it is not open-source. > This project is MIT-licensed so anyone can use, extend and contribute. ## What it does Exposes Kinaxis Maestro REST & Bulk APIs as MCP tools that AI agents (Copilot, Claude, etc.) can call to: | Tool | Description | |------|-------------| | `health_check` | Verify connectivity and auth | | `list_tables` | Discover all 800+ Kinaxis tables | | `query_table` | Run a query against any table | | `fetch_query_results` | Paginated row retrieval | | `close_query` | Release query resources | | `bulk_export` | Export up to 500k rows per batch | | `bulk_upload_init` | Start a bulk data upload | | `bulk_upload_data` | Push rows to an upload | | `bulk_upload_complete` | Finalize upload & trigger update | | `trigger_data_update` | Trigger a data update | | `data_update_status` | Check data update progress | | `run_script` | Execute Maestro scripts (sync) | | `run_script_async` | Execute scripts asynchronously | | `run_workflow` | Run Maestro workflows | | `workflow_status` | Check workflow status | | `open_workbook` | Open a workbook for data retrieval | | `read_worksheet` | Read worksheet data | | `import_worksheet` | Import data via workbook | ## Quick start ### Prerequisites - Node.js >= 18 - A Kinaxis Maestro instance with REST API access - OAuth2 client credentials **or** a web service user (Basic Auth) ### Install ```bash git clone https://github.com/federicociro/kinaxis-mcp-server.git cd kinaxis-mcp-server npm install cp .env.example .env # Edit .env with your Kinaxis credentials ``` ### Configure Edit `.env`: ```env KINAXIS_BASE_URL=https://your-region.kinaxis.net/YOUR_INSTANCE KINAXIS_CLIENT_ID=your_oauth2_client_id KINAXIS_CLIENT_SECRET=your_oauth2_client_secret KINAXIS_DEFAULT_SCENARIO=Baseline KINAXIS_DEFAULT_SCOPE=Public ``` Or use Basic Auth: ```env KINAXIS_BASE_URL=https://your-region.kinaxis.net/YOUR_INSTANCE KINAXIS_USERNAME=your_ws_user KINAXIS_PASSWORD=your_ws_password ``` ### Add to your MCP client #### VS Code (Copilot) Add to `.vscode/mcp.json`: ```json { "servers": { "kinaxis": { "command": "node", "args": ["src/index.js"], "cwd": "/path/to/kinaxis-mcp-server", "env": { "KINAXIS_BASE_URL": "https://your-region.kinaxis.net/YOUR_INSTANCE", "KINAXIS_CLIENT_ID": "your_client_id", "KINAXIS_CLIENT_SECRET": "your_client_secret" } } } } ``` #### Claude Desktop Add to `claude_desktop_config.json`: ```json { "mcpServers": { "kinaxis": { "command": "node", "args": ["/path/to/kinaxis-mcp-server/src/index.js"], "env": { "KINAXIS_BASE_URL": "https://your-region.kinaxis.net/YOUR_INSTANCE", "KINAXIS_CLIENT_ID": "your_client_id", "KINAXIS_CLIENT_SECRET": "your_client_secret" } } } } ``` ### Run standalone ```bash npm start ``` ## Authentication The server supports two authentication methods: 1. **OAuth2 Client Credentials** (recommended): Register a client in Kinaxis Admin, provide `KINAXIS_CLIENT_ID` and `KINAXIS_CLIENT_SECRET`. Tokens are refreshed automatically. 2. **Basic Auth**: Use a Kinaxis web service user with `KINAXIS_USERNAME` and `KINAXIS_PASSWORD`. Simpler setup but credentials are sent with every request. ## Kinaxis API reference This server implements endpoints from the [Maestro Web Services Guide](https://www.kinaxis.com) (REST & Bulk APIs). Place your copy of the Kinaxis documentation in the `docs/` folder (gitignored) for local reference. ## Contributing PRs welcome! Please: 1. Fork the repo 2. Create a feature branch (`feat/my-feature`) 3. Use [Conventional Commits](https://www.conventionalcommits.org/) 4. Open a PR ## License [MIT](LICENSE) Federico Berrone, 2026