Train and optimize AI agents with reinforcement learning from agent trajectories
Key Features
Single Agent Self-Hosted
ART (Agent RL Trainer) Details
| Organization | OpenPipe |
| Organization Type | Company |
| Category | Framework |
| Subcategory | Training |
| Deployment | Self-Hosted |
| Primary Language | Python |
| License | Apache-2.0 |
| Commercial Use | Permissive |
| GitHub Stars | 9,000 |
| GitHub Forks | 774 |
| Maturity | Beta |
| Pricing Model | Free |
| Self-Hosted Free | Yes |
| Cost Model | free + LLM costs |
| Community Size | Large |
| Community Activity | Active |
| Sentiment | Positive |
| GPU Required | No |
| Research Date | 2026-03-24 |
Original data from HuggingFace, OpenCompass and various public git repos.
Check out Ag3ntum — our secure, self-hosted AI agent for server management.
Release v20260328a