markuswondrak

markuswondrak / AgentMux

Public

Deterministic multi-agent pipeline for end-to-end software development, orchestrating CLI-based AI tools (e.g. Gemini, Claude, Codex) through tmux-controlled sessions.

12
0
100% credibility
Found Mar 23, 2026 at 12 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

AgentMux coordinates a team of AI agents to turn feature descriptions into reviewed, committed code using existing local AI chat tools.

How It Works

1
📖 Discover AgentMux

You find this helpful tool that lets a team of AI assistants build software features for you, just from a simple description.

2
🛠️ Get everything ready

You quickly set up the tool on your computer alongside your usual AI chat programs, no extra accounts needed.

3
🚀 Describe your feature

You type a plain English request like 'Add a login button' or link a GitHub task, and launch the team with one go.

4
👥 Watch the team collaborate

A live window opens where the planner outlines steps, the coder builds it, and the reviewer checks quality in real time.

5
Reviewer feedback time
✅
Looks perfect

Everything passes, move to wrap up.

🔧
Needs a fix

Coder makes quick changes and tries review again.

6
👍 Give final okay

You peek at the finished code, confirm it's good, and select what to save.

🎉 Feature delivered

Your new code is safely added to the project, a ready-to-review update is created, and the task is complete.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 12 to 12 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is AgentMux?

AgentMux is a Python-based deterministic multi-agent pipeline for end-to-end software development. It orchestrates CLI-based AI tools like Claude, Codex, and Gemini through tmux-controlled sessions, turning a feature prompt into planned, coded, reviewed, and committed changes. Run `python3 pipeline.py "Add rate limiting"` to kick off the workflow, with support for GitHub issues, resume, and optional product manager refinement.

Why is it gaining traction?

It stands out by reusing your existing CLI tool subscriptions—no new API keys or per-token fees—while mixing providers per role (e.g., Claude for architect, Codex for coder). The static, deterministic pipeline ensures reproducible runs, with live tmux monitoring of agents and automatic PR creation. Developers dig the hands-off automation from prompt to git commit.

Who should use this?

Indie hackers or solo backend devs prototyping features from GitHub issues. Small teams needing quick, structured code gen without vendor lock-in. Python CLI AI users wanting a multi-agent orchestrator for deterministic development pipelines.

Verdict

Try it if you're experimenting with CLI-based agents—solid docs and quickstart make it accessible despite 12 stars and 1.0% credibility score. Still early (low maturity, no tests visible), but promising for agentmux-style workflows; watch for stability as it grows.

(178 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.