multica-ai/multica

Multica is an open-source workspace that assigns coding tasks to AI agents like Claude Code, Codex, Cursor, and others as if they were teammates.

The tool solves the fragmentation problem of managing multiple AI coding agents across separate terminal sessions and contexts. Instead of juggling different agents in isolation, each losing context between sessions, Multica centralizes them in a shared workspace. When you assign an issue to an agent, it picks up the work autonomously, executes code on a runtime you control, reports progress through comments, and submits results for human review. The intent, execution, decisions, and resulting diff all stay connected to the same issue, eliminating the need to repeatedly explain context or reconstruct what happened.

Multica suits teams that want to leverage multiple coding agents without vendor lock-in or the overhead of managing each one separately. It works with twenty different agent CLIs, so you can use Claude Code, Codex, Cursor, Copilot, Kimi, and others together rather than choosing one. The tool supports organizing agents into squads with human teammates, creating reusable skills from solved problems, and running automated workflows like standups and audits on a schedule. You can self-host it on your own infrastructure, keeping code execution on machines you control, or use the cloud version. The workspace includes chat for ad-hoc questions and projects for grouping related work with attached repositories and documentation.

The project shows strong adoption signals: almost all open issues come from external users rather than the core team, indicating a substantial base of real-world users. Maintainers respond to new issues and pull requests within a day. Work in the issue tracker centers on bug reports, enhancement requests, and user questions, reflecting active engagement with the community.