uzi

New

Uzi is a developer-oriented CLI tool for orchestrating many AI coding agents in parallel. It creates isolated Git worktrees for each agent, manages a tmux session per agent, and brings up development servers on distinct ports so multiple agents can run, test, and iterate on code concurrently. Users configure projects with a uzi.yaml file specifying a devCommand and a portRange, and then spawn agent sessions using the uzi prompt command with agent types and counts. The README shows workflows for monitoring sessions, broadcasting instructions, automatically confirming prompts, and checkpointing agent changes back into the main branch. Prerequisites listed are Git, tmux, Go for installation, and any AI tool of choice such as claude or codex. The tool is intended for developers who want to run coordinated coding agents against a codebase and manage their outputs and integration.

Stars

427

App URL

https://github.com/devflowinc/uzi

Github Repository

https://github.com/devflowinc/uzi/blob/main/README.md

Features

Uzi provides parallel agent orchestration, automatic Git worktree management for isolated development, and tmux session management to keep each agent separate. It automates development server setup with port allocation based on a configurable portRange and a devCommand that runs install and dev steps per agent. The CLI includes real-time monitoring of agent status via uzi ls with a watch mode, automatic handling of prompt confirmations through uzi auto, and messaging to all agents with uzi broadcast. It offers lifecycle commands to spawn agents with uzi prompt, execute commands across agents with uzi run, terminate sessions with uzi kill, checkpoint and merge agent work using uzi checkpoint, and reset local Uzi data. Configuration is driven by uzi.yaml and supports multiple AI tool types and random agent naming.

Use Cases

Uzi speeds up experimentation and parallel development by letting multiple AI coding agents work on the same project in isolated worktrees, reducing merge conflicts and dependency collisions. Per-agent tmux sessions and automatic dev servers let agents run and preview changes concurrently on different ports, enabling faster feedback loops and testing. The monitoring and broadcast features help coordinate instructions and ensure consistent directives across agents. Automatic confirmation handling reduces manual intervention when agents prompt for trust or continuation. Checkpointing enables safe commits and rebasing of agent changes into the main branch, making it practical to incorporate useful outputs. Overall, Uzi helps developers prototype, refactor, and test code changes generated by many agents while maintaining repository hygiene and developer control.

uzi

Basic Information

Links

App Details

Categories

Similar Listings

virtual lab

mnemo

EdgeChains

RAG Agents Accelerator

LLM Zero to Hundred

Featured Listings

Terry Bison Ranch

The Singapore Flyer

Tags

More Filters

uzi