# Braintrust

Tool · [Standard tools and platforms](https://ecosystem.samcarlton.com/tools)

Also called: Braintrust AI, braintrust.dev

A platform for evaluating and [monitoring](https://ecosystem.samcarlton.com/observability) AI products. [Teams](https://ecosystem.samcarlton.com/microsoft-teams) record what their AI does, score it against tests and compare changes.

**For example:** A team uses Braintrust to compare two [prompts](https://ecosystem.samcarlton.com/prompt) on the same 300 real questions before switching.

Official site: <https://www.braintrust.dev>

## Connects to (5)

- [Evals](https://ecosystem.samcarlton.com/evals): Tests that check whether the outputs of a model or an AI product meet the criteria its builders set.
- [Testing](https://ecosystem.samcarlton.com/testing): Checking that software does what it should, before your users find the problems for you.
- [Production](https://ecosystem.samcarlton.com/production): The live version of your app that real customers use.
- [promptfoo](https://ecosystem.samcarlton.com/promptfoo): An open-source tool for testing prompts and AI apps.
- [Ground truth](https://ecosystem.samcarlton.com/ground-truth): The verified, correct answers or facts that an AI system's output is checked against.

Map view: https://ecosystem.samcarlton.com/braintrust