Braintrust

Remote Braintrust MCP endpoint that lets an assistant query an AI evaluation and observability platform's experiments, logs, datasets and prompts.

What you can do

• Query experiments and their results in the Braintrust evaluation platform
• Inspect logs and traces from production AI observability
• Read and work with datasets and prompts stored in the platform
• No tool list is published, so the exact operations available are not documented here

When to use it

Useful when your team already runs AI evaluations and observability in Braintrust and you want an assistant to look up experiment results, logs, datasets or prompts while debugging or reviewing model changes.

Requirements

A remote streamable HTTP endpoint at the Braintrust API host, with the region chosen in the address. The manifest declares no authentication, environment variables or headers, so how access is authorized is not stated; a Braintrust account and network access are presumably needed.