multica: Coordinate AI Coding Agents and Human Teams

Overview

Multica is a workspace where teams assign issues to AI coding agents as they would to teammates. Built on a Go backend, it connects a web app, desktop and mobile clients, and a local daemon that runs agent CLIs on machines you control. Agents report progress, raise blockers, and return work for human review, keeping the task, run, and changes connected.

It is aimed at teams already using coding agents who want a shared way to coordinate them with people instead of managing separate terminal sessions. You can use Multica through its cloud service or self-host it, and connect it to GitHub, GitLab, Gitea, or Forgejo.

Key Features

• Assign issues to configured agents and track work through a shared board.
• Run agent CLIs on connected computers, including your own laptop or cloud machine.
• Review execution logs containing tool calls, commands, and errors.
• Keep completed work behind a human review gate before it reaches the main branch.
• Group work into projects and give agents relevant repository and document context.
• Schedule recurring tasks such as standups, audits, and reports with autopilots.
• Connect Git hosting services and team chat channels, with access controls for workspaces and agents.
• Use the workspace through web, desktop, mobile, CLI, and API surfaces.

Use Cases

• A software team can assign routine issues to coding agents and review their changes in the same workflow as human contributions.
• A developer using several agent CLIs can coordinate their tasks and inspect what each run did without relying on separate terminal tabs.
• A team with its own infrastructure can run agents beside its code and self-host the workspace rather than use Multica Cloud.
• A team can schedule recurring repository checks or reports and review the results when an agent flags a decision or blocker.

Getting Started

For cloud access, sign up at multica.ai. For self-hosting, the README provides this installer command:

curl -fsSL https://raw.githubusercontent.com/multica-ai/multica/main/scripts/install.sh | bash -s -- --with-server
multica setup self-host

You will also need at least one supported agent CLI installed and signed in on the machine that runs agents. See the README and self-hosting guide for setup details.

License in Plain Words

No standard license was detected. Without a license, the default copyright rules apply: you may read the code, but not reuse it. Check the repository or ask the authors.

Alternatives

• AgentTeams: AgentTeams coordinates agents in shared Matrix rooms, while multica centers work on issues assigned to coding agents.
• codeg: Codeg organizes agent conversations, sessions, and code changes around ACP, while multica focuses on issue assignment and review.
• vicoa: Vicoa offers agent management across desktop, web, mobile, and CLI, while multica centers collaboration on assigned issues and execution progress.

Considerations

• The project is active, with regular commits and releases, but its measured history is still relatively young. High activity is not a guarantee of long-term stability.
• Self-hosting involves a Go backend, PostgreSQL, and Docker. The README describes Docker Compose installation; Helm is not listed as a supported requirement option.
• Multica does not supply agent models or CLIs. You must install and authenticate a supported CLI on the runtime machine.
• The repository describes its license as the Multica License, based on Apache 2.0 with additional conditions. Review the exact license terms, especially for hosted services, commercial embedding, and branding.
• The README says the iOS app builds from source but is not yet available on the App Store.