OpenExp
Save, reuse, and share your AI agent's successful trajectories.
OpenExp is an open-source framework that lets users record, share, and replay successful AI agent sessions as structured 'exps' — four-file bundles capturing prompts, reasoning, skills, scripts, and decisions. These exps can be installed locally into Claude's skills directory so future agents can follow proven trajectories instead of reasoning from scratch. The project is MIT-licensed, local-first, and requires no cloud or API keys.
Users record an AI session trace into a four-file exp bundle and install it into their local Claude skills directory, allowing future agent runs to reference and replay the trajectory that led to a verified result.
developers and teams using AI agents like Claude who want to capture and replay successful workflows
Background.
- Status
- launched
- Business model
- open-source
Founders
- ivan-pasichnykgithub
Similar projects.
Editorial take on the space this project sits in — momentum signals, adjacent moves, our call on whether the wedge is real. Get pinged when we publish a new read or when the landscape shifts.
Have a take on this space?
Tell us what you’d build differently, where you think the incumbents miss, or what we’ve gotten wrong about this project. Comments + reactions are coming soon.