yorgai/org2 — explained in plain English
Analysis updated 2026-07-03 · repo last pushed 2026-07-03
Review what an AI agent did overnight on a feature by watching the replay.
Debug and audit AI coding sessions step by step.
Manage multiple AI agents working on different tasks.
Use AI for coding with full transparency and local data control.
| yorgai/org2 | openkursar/hello-halo | f/git-rewrite-commits | |
|---|---|---|---|
| Stars | 1,428 | 1,423 | 1,411 |
| Language | TypeScript | TypeScript | TypeScript |
| Last pushed | 2026-07-03 | — | — |
| Maintenance | Active | — | — |
| Setup difficulty | moderate | easy | easy |
| Complexity | 3/5 | 3/5 | 2/5 |
| Audience | developer | pm founder | developer |
Figures from each repo's GitHub metadata at analysis time.
Desktop app is under 100MB but optional features like browser automation and macOS computer control require separate helper programs.
ORG-2 is an open-source AI coding assistant that runs on your desktop and helps you write, review, and manage code. Think of it as an alternative to tools like Cursor or Claude Code, but with a sharper focus on letting you see exactly what the AI did, why it did it, and how to hold it accountable for its work. Instead of treating the AI like a black box that just spits out code, this project records everything the AI agent does during a session. You can replay those actions later, step by step, to audit what happened, debug issues, or share with teammates. Agents also keep memory across sessions, so they remember context from past work rather than starting fresh every time. The software runs locally on your machine using a Rust-based engine, meaning your data and execution stay on your device rather than being sent to a remote server. The tool is built for developers, team leads, and technical founders who want to use AI agents for coding but also need transparency and control. For example, if you are a team lead reviewing what an AI agent did overnight on a feature, you can watch the replay rather than guessing. If you are managing multiple agents on different tasks, the project is working toward org-level coordination features like issue tracking and shared accountability, so humans and agents can collaborate around common goals. What makes this project notable is its philosophy. Rather than positioning AI as a simple output machine, it treats agents as persistent, observable colleagues in a structured workflow. The desktop app is lightweight at under 100MB and supports a wide range of built-in tools, including GUI, terminal, Git, browser automation, and database tooling. Some advanced features, like computer control on macOS and browser automation, rely on optional helper programs you can download separately. The project is actively developed, available for macOS, Windows, and Linux, and has an active community on Discord and WeChat. It is licensed under AGPL-3.0, meaning it is fully open-source with requirements around sharing modifications if you distribute or host it publicly.
A desktop AI coding assistant that records everything it does so you can replay, audit, and manage its work. Runs locally with session memory and transparency controls.
Mainly TypeScript. The stack also includes TypeScript, Rust.
Active — commit in last 30 days (last push 2026-07-03).
Fully open-source, but if you distribute or host modified versions publicly, you must share your modifications under the same license.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
double-check against the repo, no cap.