- Home
- MCP servers
- Coda
Coda
- javascript
0
GitHub Stars
javascript
Language
4 months ago
First Indexed
2 months ago
Catalog Refreshed
Documentation & install
Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.
You set up a Coda MCP Server to expose a live, streamable MCP endpoint for your client tools, with a fallback legacy SSE endpoint. This lets you publish data sources and actions so MCP clients can connect, stream updates, and automate workflows. The server can be deployed on Koyeb for cloud hosting or run locally for testing and development.
How to use
You connect MCP clients to the primary streamable endpoint at /mcp to receive real-time updates. There is also a legacy SSE endpoint at /sse for compatibility with older clients. The server includes a health check endpoint at /health to verify availability. When integrating with LangSmith or other tooling, you will configure an MCP server pointing to the deployed mcp URL and, if needed, provide authentication via static headers.
How to install
Prerequisites you need before installing: Node.js installed on your machine, and a package manager such as pnpm.
Step 1: Prepare your environment
- Ensure you have Node.js installed
- Ensure you have pnpm installed
- Have your Coda API key ready (you will use API_KEY or CODA_API_KEY) and know the port you want to expose (default is 3000)
Step 2: Install and build locally
- Clone or copy the repository contents to your working directory
- Run the following commands to install dependencies and build the server:
pnpm install
pnpm build
echo "API_KEY=your-coda-key" > .env
PORT=3000 pnpm start:http
Step 3: Verify locally
- Health check: http://localhost:3000/health
- MCP endpoint: http://localhost:3000/mcp
Additional configuration and deployment
You can deploy the MCP server to Koyeb and expose it publicly. Typical setup includes using a web service, linking your GitHub repo, and configuring environment variables for authentication and port.
Koyeb deployment steps you will follow (summary):
- Create a new Web service
- Connect your GitHub repository and set the main branch
- Use the builder/preset as the build method
- Allocate a Free / CPU Eco instance
- Add environment variables: API_KEY and PORT
- Configure port 3000 with HTTP protocol and enable public HTTPS access
- Deploy and access the public URL, health endpoint, and MCP endpoint
LangSmith integration for the MCP server requires adding an MCP server with a URL corresponding to your deployed mcp endpoint and selecting Static Headers authentication if you need to provide a header-based token. Use the dedicated MCP URL path /mcp and avoid using the root or /sse path in the builder configuration.
Available tools
health
Health check endpoint at /health to verify the server is running and accessible.
mcp
Streamable MCP endpoint at /mcp used by MCP clients to receive real-time updates.
sse
Legacy SSE endpoint at /sse for compatibility with older clients.
agent_builder
LangSmith Agent Builder integration to connect the MCP server for automated testing and monitoring.
deploy_koyeb
Koyeb deployment workflow to publish the MCP server to a public URL with configured ports and environment variables.