Codeg is a multi-agent coding workspace that aggregates sessions from multiple AI coding agents into a single collaborative environment.
The tool solves the fragmentation problem of working with different AI coding agents by centralizing them in one searchable workspace where a main agent can delegate tasks to sub-agents of other types. It supports agents including Claude Code, Codex, OpenCode, Pi, Grok Build, and others, with fifteen agents built in and the ability to register additional ACP-compatible agents. Work can be organized on a to-do board where each task runs in its own branch unattended, awaiting review before integration.
Codeg suits developers who regularly use multiple AI coding agents and want to coordinate their work rather than switching between separate interfaces. The tool is particularly valuable for teams or individuals managing complex coding tasks that benefit from delegating to specialized agents. It runs flexibly as a desktop application, a self-hosted server, or a Docker container, with native iOS and Android clients for remote access.
The project shows consistent development activity with regular commits across its codebase. The maintainer actively responds to issues and pull requests, indicating ongoing engagement with the user community. Documentation is comprehensive and maintained alongside the code, suggesting a commitment to usability. The tool receives periodic updates that address both bug fixes and feature enhancements, demonstrating sustained investment in its stability and capabilities.
