Multica
Open-source Kanban platform for human-AI agent teamwork. Assign issues to AI agents powered by Claude Code, Codex, etc., and let them execute tasks autonomously while your team focuses on oversight.
Tool overview
Multica is an open-source platform for orchestrating AI agents as first-class team members. It provides a Kanban-style task board where issues can be assigned to both humans and agents. Agents run on a local daemon, calling tools like Claude Code or Codex, with API keys and code kept locally. It supports desktop, cloud SaaS, and self-hosted deployment.
Pros: Real human-AI task collaboration, reusable skill accumulation, automated retry on failure, and strong privacy. Cons: Steep learning curve; agent teams require meticulous prompt engineering; the 7-state workflow may feel rigid; Kanban UI less natural than chat for some. Suits dev teams comfortable with AI tooling seeking to manage multiple agents and compound workflow knowledge. Not ideal for teams wanting zero-config or highly customizable workflows.