Enables AI systems to query and analyze OpenTelemetry traces and metrics through Logfire's API, providing tools for finding exceptions, investigating errors, and running custom SQL queries against observability data with automatic authentication.
Log in to Pydantic Logfire
Navigate to Your Project's "Read Tokens" Section
- Ensure you are in the correct project for which you want to give MCP server access.
- Under project settings, find the "Read Tokens" section.
Create a New Read Token
- Click the button to create a new read token.
- (Optional) Give the token a name or description to help you remember its purpose.
- Copy the newly generated token. You will need this token for the next step.
Fill in the FastMCP Connection Interface
- Click the "Install Now" button for the Pydantic Logfire MCP Server integration.
- In the FastMCP connection interface, fill in the required ENV value:
- For
LOGFIRE_READ_TOKEN, paste the read token you copied from Pydantic Logfire.
- (Optional) If you need to specify a different Logfire API base URL (for non-default regions or self-hosted instances), fill in the
LOGFIRE_BASE_URL as provided by your API host.
Save and Finish Installation
- Submit the configuration in the FastMCP interface.
- The Pydantic Logfire MCP server is now connected and can analyze your OpenTelemetry traces and metrics.
Note: Each read token is specific to a project. If you want to connect multiple projects, repeat these steps for each project.
