> Best of luck with this but I think with so many open source agent managers cropping up

What’s the top 5 (or any N) that come to mind:

A) GUI based

B) terminal based

C) web based?

Like, not just personal projects but something with a bit of a community around it? I remember Conductor from a bit ago (seems only Mac is supported) and a few other HN posts but all of those seemed smaller and more barebones. Oh I guess OpenCode also has a desktop and web version, but it never worked well for me (and I need something that can just use headless Claude Code instances).

Asking because I just use Claude Code desktop for organizing my sessions and am a bit behind in that regard - if there are indeed many options that others can vouch for somewhat, I’d love to hear about them!

Edit: apparently there is Cmux (Mac only), T3 Code (very new), Agent Orchestrator (tries to be a weird kanban board), Agor (tries to be a weird canvas board) and Claude Squad (TUI only), but none of those are quite what I'm looking for. If there's all that many options, I might have missed most of them - since Baton or OpenCode (a revisit of it) seem more like what I'd be looking for, maybe Conductor if not Mac only.

I appreciate the feedback!

I have two claude code subscriptions: a team plan through my employer and I'm paying for the $200/month plan outside of that.

Trading $200/month of my money for the ability to build all of the things I've been thinking about for years is a great trade for me. I've built more things for fun/potential profit in the last year than I did in the previous decade combined.

And of course, one of the things I've built is a version of what OP made that works exactly how I want it to work. :)

I build my own products and services and the effective ROI for paying for a more or less unlimited max Claude Code plan is fairly ridiculously positive.

This uses the CLIs so its using subscription pricing, not token pricing

VC funding + spending more money on Claude instead of hiring more engineers

200 dollars a month goes a long way with claude code

People are just playing around with parallel agents because it looks cool on Twitter. In real prod, 90% of your time isn't spent typing lines of code, it's spent trying to figure out implicit business requirements and debugging undocumented legacy spaghetti. Agents sitting in isolated worktrees are completely useless here - they'll just rapidly and in parallel write code that completely fails to solve the actual business problem

You could zoom out a bit and rephrase the question.

Your great-aunt Ida died and left you a consulting team of ten pretty good software engineers. The team's contracts all just ended, so starting tomorrow they'll be idle. Ida said you must run the business for at least two years (fortunately, overhead is already paid for), or forfeit your share of the inheritance. After that you can keep going or liquidate it.

What do you do?

You’re seeing that primarily because it’s what people can show off easily. Side projects they do for fun.

I use all of this stuff daily at work. Normally, I’m working on 2 to 4 features in parallel (so worktrees). This might not be simultaneously, but it’s at least across days or weeks.

Skills, agents, tasks, etc are really about creating repeatability in certain parts of my workflow without needing to be hands on.

Some changes within pre-existing codebases (oh hey, we need feature or mechanism X).

An entirely new internal system with LLM code review, DB migration tracking, time tracking, standups and Teams integration.

A new system that trains neural nets to recognize crops based on Sentinel-2 satellite data (the neural net works okay, mowing and ploughing is harder with a mostly heuristic approach since I don't really have labels).

A new system to migrate somewhere between 1000-2000 forms between proprietary solutions where a team of people have spent a year with limited progress, whereas I'm generating the codegen tool that does most of the work, with the remainder being left up to AI.

A new project linting tool with Go + goja to allow writing rules for validating project stuff in ECMAScript, a bit like ESLint just stack agnostic and can be deployed as a 10 MB executable, to also control stuff like architecture and project conventions that the other tools aren't really geared towards.

Also wrote an OpenAI/Ollama/Claude proxy that allows using on-prem models running on another server through Ollama/llama.cpp and also using AWS Bedrock models when permissions are configured.

Also a bunch of Ansible configuration for stuff like a self-hosted Sentry instance, debugging that piece of shit would be so hard and annoying without something that I can throw logs at (because for some reason they think that having 70 containers running for what should amount to one piece of software is okay).

Also wrote a personal tool that lets me use VLM and Whisper and PySceneDetect and some other stuff to produce EDL so I can take a 3 hour long video and cut it down to 1 hour with LLMs using the transcripts/timestamps (aligned with words, so not too many awkward cuts) that I can then import into DaVinci Resolve for further editing.

Also migrated the apps I host from Contabo VPSes to Proxmox VMs (Hetzner dedicated server from the auction) and went from Docker Swarm + Portainer to pure Docker Compose, also moved from Drone CI to Woodpecker CI and also got rid of the old deprecated Bitnami container images.

Also migrated my homepage from an ancient Ruby and Rails version to more modern ones.

Also wrote a few scripts to replace YOURLS with just an Apache install, the config for which I can automatically append new shortened links to.

I don't even need worktrees or custom skills for most of this, just Claude Code and a subscription, since paying per token would make me go broke.

Yeah perfect example, the main thing I _would_ use multiple agents on is optimizing/benchmarking code, but for that you specifically can't use worktree, you need one agent per machine or they'll taint each other's benchmarks

Appreciate the feedback! Looking into it

Fair but FWIW I love a GUI and I’m not gonna complain if everyone and their mother want to offer options

Let a thousand vibecoded flowers bloom

The main difference is that Baton is agent-agnostic and terminal-native. It doesn't add a GUI on top of Claude Code or Codex, it builds around the terminal itself, so you run whatever agent CLI you want natively, but with convenient shortcuts for launching them. Which is a nicer experience in my view, but people have different views on this.

Baton is also more git-aware. Instead of just showing raw diff line counts, you see commits ahead and behind your target branch, so you can tell at a glance how far each workspace has diverged and shortcuts for resolving it in the matter you want.

One thing I think is unique is the built-in MCP server. It lets agents spawn new workspaces programmatically, so you use an agent to launch agents in new isolated workspaces.

Would be curious if it is more polished than Conductor. Memory leaks and random bugs seem to crop up in Conductor far too often.

If nothing else, I see that Conductor is currently Mac only.

The underlying git worktree still lives on your disk until you delete it. So its not harder than starting a terminal with claude --continue, or codex resume --last inside the git worktree, depending on what agent the user used.

I have tried to provide after best ability, but have only been testing them on vm's on my mac! So be aware. I labeled them Beta due to this. But most features should work fine, probably better on linux than windows.

You can open a directory also as a workspace, it just wont have git stats and git shortcuts.