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. Prompt Optimization
6. Judges and Instruction Following
7. Tool Use and Search
8. Multimodal Environments
9. Coding Agents and Sandboxes
10. Synthetic Agent Environments
11. Custom Harnesses
12. Best Practices
13. Legacy Environments — reference only; not part of the main path

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.

This is a local map of the cookbook workspace.

• guides - ordered walkthroughs.
• configs - runnable Prime eval, GEPA, and training configs.
• environments - local environment packages used by guides.
• reference - thin pointers only; do not duplicate managed guidance here.