baseline

Baseline MCP Server provides access to the Web Platform Dashboard's API to retrieve the Baseline (compatibility) status of Web Platform features. It exposes a queryable MCP interface that returns information such as feature baseline status (widely, newly, limited, no_data), browser support details (versions and implementation dates), usage data for features, and allowlist/exclude controls for specific browsers. You can connect to this MCP server through an MCP client (for example Claude Desktop or Visual Studio Code integrations) and request data about Web Platform features, which the server fetches from the Web Status API and returns to your client.

To use it, run the server in Docker (as described in the installation guide) and configure your MCP client to point at the running container. The server supports interacting with the Web Platform Dashboard to look up features, their baseline statuses, and implementation details. You can use the client’s settings to filter out certain browsers (chrome, edge, firefox, safari) if you want to narrow results, and you can integrate the data with AI-assisted workflows via the MCP protocol to coordinate with other AI models in your pipeline.

Prerequisites:\n- Docker installed on your machine.\n- Optional: An MCP client (e.g., Claude Desktop or VS Code MCP extension) configured to talk to a running MCP server.\n\nInstall steps:\n1) Build the Docker image (as described in the README):\nbash\ndocker build -t baseline-mcp-server .\n\n2) Run the MCP server container (keep it running and accessible to your MCP client):\nbash\ndocker run -i baseline-mcp-server:latest\n\n3) Configure your MCP client to connect to the running server. For example, in Claude Desktop or VS Code settings, point to theBaseline MCP server with the appropriate command and args as shown in the README. Ensure the client is allowed network access to api.webstatus.dev if you are using the baseline data source.\n\nNotes:\n- The server relies on the Web Platform Dashboard API; ensure network access to that service is permitted.\n- If you prefer not to use Docker, you can adapt the workflow to run the server natively if you port the containerized image to your environment, but the provided instructions assume Docker usage.

Tips and considerations:\n- The server fetches Baseline status data from the Web Platform Dashboard API; expect occasional latency if the external API rate limits or experiences downtime. Consider caching results in your MCP client or network layer if you plan to run frequent queries.\n- When using Docker, you can pass environment variables to customize behavior (e.g., API endpoints, feature filters) if the image supports such overrides in a future update. Currently, the documented approach uses the default configuration with the image baseline-mcp-server:latest.\n- If you need to exclude certain browsers from searches (e.g., chrome, edge, firefox, safari), use the MCP client’s filtering capabilities as described in the Baseline MCP Server documentation.\n- The server is designed to be used with an MCP client that reads and writes through the MCP protocol; ensure your client version supports the same server interface.\n- The recommended runtime for development is Docker; however, the server is not Node.js-based, so there is no npm package name associated with this particular deployment.