codeg

codeg is a multi-agent AI coding workspace that aggregates sessions from multiple coding assistants, such as Claude Code, Codex, OpenCode, and Gemini CLI, into a single interface. Instead of switching between separate tools, developers can view and manage conversations from all agents in one place, reducing context loss and tool-hopping.

In practice, it lets you compare code generations or suggestions from different AI agents side by side, helping teams make faster, more informed decisions. The unified view is especially useful during code review or when exploring alternative implementations.

Key advantages: open-source and free; supports desktop, self-hosted server, and Docker deployment, giving flexibility and control over data. The project has gained 1,900 GitHub stars, showing early community interest. Limitations: still in early stage, so documentation and stability may be limited; requires technical know-how to deploy and maintain (Docker or server management), which can deter non‑technical users. It currently aggregates a specific set of agents, so not all tools are covered.