← Back to Discover

Multica

A workspace where you assign issues to AI coding agents as you would to teammates. It drives agent CLIs you already use on your own machines and tracks runs, cost and review.

Coding agentsTask management
Visit website ↗

What it does

Multica is a workspace for running AI coding agents the way you run a team. You file an issue and set an agent as the assignee. The agent picks the work up, runs it on a machine you control, comments as it goes, raises blockers and moves the issue to review. The intent, the run, the decisions and the diff stay attached to the same issue, and the README says nothing ships without a human approving it. It does not ship a model. A daemon on your laptop or a cloud box drives the agent command-line tools you already have installed and signed in, 26 of them listed today, including Claude Code, Codex, Cursor, GitHub Copilot CLI, OpenCode, Kimi and Qwen Code. Each agent gets a name, a provider and a runtime and shows up on the board like anyone else. Squads group agents and people, skills turn a solved problem into a reusable playbook, and autopilots run recurring jobs such as audits and reports on a schedule. For oversight there is an execution log that replays every tool call, command and error, the ability to reply to a running agent, per-run and per-issue cost analytics, review gates, an inbox, and automatic retries and timeouts. You can self-host with Docker Compose or Helm, connect GitHub, GitLab, Gitea or Forgejo, set owner, admin and member roles, and trigger work from Slack, Feishu/Lark, DingTalk, WeCom or Telegram. A hosted cloud version, a desktop app for macOS, Windows and Linux, a CLI and an API are also described. The license is Apache 2.0 with additional conditions on hosted services, commercial embedding and branding, and the README calls the project source-available. Self-hosted servers send one anonymous daily usage snapshot unless you set DO_NOT_TRACK=1.

Capabilities

  • Assign issues to AI coding agents as if they were teammates
  • Works with 26 agent CLIs, including Claude Code, Codex and Cursor
  • Runs on your own machine through a local daemon
  • Execution logs, cost analytics and review gates
  • Reusable skills, squads and scheduled autopilots
  • Self-host with Docker Compose or Helm

When to use it

  • You run several coding agents and want one board to assign and track their work
  • You want agents to work on your own machines so code stays there
  • You need to see what each agent run did and what it cost
  • You want to self-host the whole stack alongside your own Git host