Version Control for AI Agents
Git for agent sessions lets you swap models mid-task and branch reasoning trees.
Version-Control for AI coding agents.
Git-style blame for AI agents when nothing else tracks prompt-to-line causality.
Developers using AI coding agents like Claude Code
Git · Cursor · Warp
vcs.
i find myself struggling with questions that agents can't answer like "why did you do it?", "when did u delete this folder? why?", etc. or trying to /rewind (after a /compact...) or basically `bisect` to find when and why something was done by the agent in the current / previous session.
just like git did for code, i think we are the same core capabilities with ai agents
so...
i developed an open source solution for that (currently supporting claude code)
would love to get feedback, contribution or maybe other ideas or solutions you find for those problems.
Git for agent sessions lets you swap models mid-task and branch reasoning trees.
Git branches for agent memory with time-travel rollback via MatrixOne CoW engine.
CRDT-based VCS for AI swarms where Git's locking model fails completely.
Git for agent cognition—clever framework, but no working implementation yet.
Git-style branching for docs that agents can actually read via MCP.
Git for agent memory beats Letta's file-level versioning with zero-copy branching.