Einsia/agent-git

Code is cheap, show me the talk: version control for agent sessions

What it solves

Agent sessions are typically stored as temporary, volatile files that are easily lost or overwritten. This tool provides a reliable version control system for these sessions, allowing you to save, resume, and share specific points in an agent's history.

How it works

It acts as a version control layer on top of your agent's local JSONL session files. You can create snapshots of your work, browse history, and publish sessions to a hub. It supports common agent tools like Claude Code, Codex, OpenCode, and Cursor, and provides a terminal interface to manage these sessions, fork them, or continue them exactly where you left off.

Who it’s for

Developers and power users who work with AI coding agents and need a way to track, save, and collaborate on specific agent-driven sessions without losing progress.

Highlights

  • Lossless Versioning: Turn ephemeral agent sessions into persistent, resumable, and shareable history.
  • Cross-Tool Compatibility: Works with popular agents like Claude Code, Cursor, and OpenCode.
  • Terminal Interface: Manage sessions, logs, and publishing directly from your command line.
  • Easy Setup: Available as a prebuilt binary via npm or source build, with no Rust toolchain required for standard users.
  • Built-in Safety: Includes local secret filtering and repository secret management to keep your sessions secure.

Related

  • Dispatch
    Version Control for AI Agents: Beyond the Git Commit

    An exploration of Regent VCS, a tool designed to track AI agent intent and session history, and the debate over whether agents need their own version control or can simply use Git.

  • Project
    Dicklesworthstone/coding_agent_session_search

    A terminal and JSON CLI tool, written in Rust, that indexes the local conversation history of dozens of different AI coding agents into a single searchable database. It provides high-speed full-text search, optional local semantic search, and a 'robot mode' API so AI agents themselves can query past sessions. It emphasizes privacy (everything local), robustness (atomic index updates, quarantine of corrupt data), and performance (sub-60ms searches).

  • Project
    ling-kong-ran/pisper

    A multi-agent application that allows users to manage AI conversations using a Git-like branching system to explore multiple ideas in parallel across desktop, mobile, and terminal.

  • Project
    computerlovetech/agr

    A package manager for AI agent skills that allows teams to install, version, and sync skills from Git repositories across multiple AI tools.

  • Project
    Ancienttwo/repo-harness

    A file-backed workflow and runtime for AI agents that replaces chat memory with repository-based state, enabling resumable sessions and token-efficient development.