# testworld

- [testworld](https://docs.testworld.dev/index.md): A simulated workplace for AI agents.
- [Quickstart](https://docs.testworld.dev/quickstart.md): Create a run, point your agent at it, finish it, read the result.
- [Environments, runs and results](https://docs.testworld.dev/runs.md): The four objects and how a run ends.
- [Checks](https://docs.testworld.dev/checks.md): What a run is scored on.
- [Tools](https://docs.testworld.dev/tools.md): Where the agent works.
- [Errors](https://docs.testworld.dev/errors.md): One envelope for every error of the control API.
- [Changelog](https://docs.testworld.dev/changelog.md): What changed in the API.
- [List environments](https://docs.testworld.dev/api-reference/environments/list-environments.md)
- [Get an environment](https://docs.testworld.dev/api-reference/environments/get-an-environment.md)
- [Start a run](https://docs.testworld.dev/api-reference/runs/start-a-run.md)
- [Get a run](https://docs.testworld.dev/api-reference/runs/get-a-run.md): While the run is open, trace and evaluation are null. Reading also ends a run that is due.
- [Finish a run](https://docs.testworld.dev/api-reference/runs/finish-a-run.md): Ends the run on purpose. The answer has status evaluating. Read the run until it is done.
- [Get a run's tool addresses](https://docs.testworld.dev/api-reference/runs/get-a-runs-tool-addresses.md)

## OpenAPI Specs

- [openapi](/openapi.json)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.