multica

Why we included this project

Most people running coding agents are still babysitting them: pasting prompts, watching terminal output, re-explaining the same context on every run. Multica is built to change that. It gives each agent a name, a provider, and a runtime, then puts them on a shared board where you assign issues the way you would to a colleague. The agent picks up the work, comments as it goes, and hands the result back for review, with the execution log, token usage, and review gates attached to the same issue. Nothing ships without a human signing off. It works with a wide range of agent CLIs and can be self-hosted on your own machines, which matters if you want code to never leave your infrastructure. The payoff is for teams that want to stop babysitting individual agent runs and start managing a mixed human-and-agent workforce.