OpenPipe
Easily apply reinforcement learning to any agent
- Category
- AI & Machine Learning
- Launched
- July 11, 2025
- Stage
- Unknown
- Founder
- kcorbitt
- Website
- openpipe.ai
OpenPipe provides a post training platform that makes it easy to apply reinforcement learning (RL) to any AI agent. Its key technology is the open source Agent Reinforcement Trainer (ART), which uses GRPO powered feedback loops to continuously optimize models from production data. The platform also supports supervised fine tuning (SFT), on premises and VPC deployment, and includes features like unified observability, automated guardrails, and compliance support (SOC 2 Type II, HIPAA, GDPR).
Discovered via
- Hacker News Show HN · · Show HN: RULER – Easily apply RL to any agent