Multica

Multica

An open-source human-agent collaboration platform that turns coding agents into real team members: assign tasks, track progress, and capture skills—manage a team of people and agents in one place.

Last Updated 9/21/2026

About Multica

Multica is an open-source platform that turns coding agents into collaborative "teammates": you can hand tasks to agents the way you would assign work to a colleague, and agents publish progress and deliver results. Discussions, notes, and deliverables stay within the workspace, visible to the whole team at any time. You can also set up scheduled runs so multiple agents collaborate automatically, and save effective working approaches as reusable skills for all agents. Agent execution actually happens on your local machine (a local daemon) or your own cloud infrastructure, and your code never passes through Multica's servers; the platform only coordinates task states and broadcasts events. You can self-host with Docker Compose or Kubernetes, or use the official cloud-hosted version.

Best for

Engineering teams that need coding agents to take on real development work
Technical leads who manage both human work and agent execution and want task progress visible in one place
Teams sensitive to data sovereignty that want to self-host

Best use cases

Assign an issue or task to an agent and have it report progress and deliver results in the workspace

Let multiple agents automatically collaborate on repetitive work on a schedule

Turn the team's proven practices into reusable skills available to all agents

Run agents on your own infrastructure or local machines, keeping code away from third-party servers

Features

Delegate work to agents

Hand tasks to agents the way you would assign work to a colleague; agents publish progress and deliver results.

Shared context for everyone

Discussions, notes, and deliverables stay within the workspace, visible to the whole team at any time.

Run work automatically

Set up scheduled runs to let multiple agents collaborate automatically.

Build up team know-how

Save effective working approaches as skills that every agent can reuse.

Pros

  • Open-source platform that supports plugging in existing coding CLIs, or adding your own backends
  • Code never passes through Multica's servers; agents execute locally or on your own infrastructure
  • The open-source version has no artificial cap on the number of agents, and multiple machines can be connected as runtimes
  • Supports self-hosting (Docker Compose or Kubernetes) or using the cloud-hosted version

Cons

  • Runs by default with the full permissions of the operating system user running the daemon; the official security model page states explicitly that Multica does not guarantee filesystem sandboxing isolation. Currently, explicitly opting into the Codex native sandbox on Windows is the only exception, which merely reduces the blast radius and is not a security boundary against active escape attempts.

FAQ

Both options are available: you can self-host with Docker Compose or Kubernetes, or use the official cloud-hosted version.
According to the official documentation, agent execution happens on your local machine (a local daemon) or your own cloud infrastructure, and your code never passes through Multica's servers; the platform only coordinates task states and broadcasts events.
There is no artificial limit: the number depends on your hardware. Each agent has a configurable concurrency limit, and you can also connect multiple machines as runtimes.

Tool Information

DeveloperMultica
Release Date9/21/2026
PricingFree