Lab Cookbook

Create, evaluate, train and deploy your own agents with Prime Intellect Lab.

1. Setup
2. Environments and Evals
3. Building Your First Environment
4. Training with RL
5. Warm Starts with SFT
6. Prompt Optimization
7. Coding Agents and Sandboxes
8. Tool Use and Search
9. Synthetic Agent Environments
10. Multimodal Environments
11. Custom Harnesses

For platform reference (accounts, secrets, Hub workflows, hosted runs, inference deployments), see Lab Configuration. For quick definitions of Lab, Verifiers, evaluation, and RL terms, see the Glossary.