Guides
Task-oriented walkthroughs, in a learning order.
Task-oriented walkthroughs, in a learning order. Every command here has been run; every output is real.
| Guide | Gets you |
|---|---|
| quickstart.md | A real leaderboard in 60 seconds, no API key |
| your-first-project.md | Your own task evaluated end to end |
| local-models.md | Models on your own machine, beside frontier APIs |
| api-keys-and-gateways.md | Several keys, custom gateways, secret references |
| custom-scorers.md | Grading only you can write |
| comparing-pipelines.md | A whole multi-agent system on the leaderboard |
| ci-integration.md | Gating a merge on evaluation results |
| continuous-evaluation.md | arena watch — catching a model that drifted since you picked it |
Source: docs/guides/README.md