Naka-pin na Tweet
prateek
4.7K posts
prateek
@agent_wrapper
Orchestrator of agent orchestrators (@aoagents) @Composio
SF Sumali Nisan 2019
780 Sinusundan5K Mga Tagasunod
@cognition @composio @aoagents composio launch post on 20th Feb: x.com/agent_wrapper/…
join us! jobs.ashbyhq.com/composio
prateek@agent_wrapper
We just open-sourced the system we use to manage 30 parallel AI coding agents per person. 40K lines of TypeScript. 3,288 tests. 17 plugins. Built in 8 days — by the agents it orchestrates. Yes, we used Agent Orchestrator to build Agent Orchestrator. Some numbers: → 500+ agent-hours in 24 human-hours (20x leverage) → 86 of 102 PRs created by AI (84%) → After Day 4, I stopped writing code entirely Spawn agents. Step away. Ship faster.
English
This is very cool. @cognition just launched an agentic orchestration layer for devins. Will have to check out to see how well it's architected.
@composio was really ahead of the game on this
We launched @aoagents 1 month ago, and it is completely open source and pluggable
We are already outspeeding AI Labs in big product launches, please join us and help us push the frontiers even more
we are agressively hiring top talent across SF and Bangalore right now
both ashby link and original launch post link in comments
Cognition@cognition
Devin can now manage a team of Devins. Devin will break down large tasks and delegate them to parallel Devins that each run in their own VM. Over time, Devin gets better at breaking down and managing tasks for your codebase. Available now for all users.
English
prateek nag-retweet
yay! @cognition in on the orchestration bandwagon. if you wanna use a completely pluggable open source coding orchestrator, use me :)
Cognition@cognition
Devin can now manage a team of Devins. Devin will break down large tasks and delegate them to parallel Devins that each run in their own VM. Over time, Devin gets better at breaking down and managing tasks for your codebase. Available now for all users.
English
@alxdrnguyen would love to see you on AO discord, we meet everyday at 10 PM IST
discord.gg/e2Zpy7SpX
English
@agent_wrapper It has been an awesome tool to use. Thanks for open sourcing this.
English
that's exactly how @aoagents should be used! And we are actively working on making the OpenClaw integration as good as possible!
TheAIDataGuy@AIDataDude
@hasantoxr I just installed this and wired it into my openclaw mission control and its sick!
English
@hasantoxr thanks for the shoutout @hasantoxr ! :)
Small correction, we just hit 4.7K stars
Also, looking for contributors here
discord.gg/UZv7JjxbwG
We meet everyday at 10 PM IST
English
🚨 BREAKING: Composio just open sourced a tool that lets you run 30 AI coding agents in parallel on your codebase.
It's called Agent Orchestrator and it's sitting at 4.1k stars for a reason.
One command spawns an agent. It gets its own git worktree, its own branch, its own PR. CI fails? Agent fixes it. Reviewer leaves comments? Agent addresses them.
Here's what it actually handles:
→ Spawns parallel Claude Code, Codex, or Aider agents
→ Isolated git worktrees so agents never step on each other
→ Auto-fixes CI failures by sending logs back to the agent
→ Auto-addresses code review comments without you touching anything
→ Real-time dashboard at localhost :3000 to supervise everything
→ Notifies you only when human judgment is actually needed
One command:
ao spawn my-project 123
That's it.
Without this, you're manually creating branches, babysitting agents, reading CI logs, forwarding review comments, and tracking which PRs are ready.
With this, you spawn and walk away.
This is the missing coordination layer every team running AI agents in production needs.
100% Open Source. MIT License. Built by Composio.
(Link in comments)
English
@ofer_shapira @aoagents @composio very cool, you should add star graphs to it!
gitshow.dev/ComposioHQ/age…
English
@agent_wrapper @aoagents @composio wow, parallel coding agents sound intense, especially on real repos i built gitshow.dev btw, it makes a page from your github
English
agent orchestrator (@aoagents) just crossed 𝟰,𝟱𝟬𝟬 ⭐ on GitHub
it started as an internal dev tooling side project at @composio. we built it to run parallel coding agents on real repos, open sourced it, and the community made it way bigger than we expected.
growth has been fully organic. no paid push.
the key distinction:
agentic orchestration != sub-agents
agentic orchestration != static workflow wiring
our value prop is the 𝗯𝗲𝘀𝘁 𝗼𝗳 𝗯𝗼𝘁𝗵 𝘄𝗼𝗿𝗹𝗱𝘀:
an 𝗮𝗴𝗲𝗻𝘁𝗶𝗰 𝗼𝗿𝗰𝗵𝗲𝘀𝘁𝗿𝗮𝘁𝗶𝗼𝗻 𝗹𝗮𝘆𝗲𝗿 𝗼𝘃𝗲𝗿 𝗱𝘂𝗿𝗮𝗯𝗹𝗲, 𝗶𝗻𝘁𝗲𝗿𝗮𝗰𝘁𝗮𝗯𝗹𝗲 𝗮𝗴𝗲𝗻𝘁 𝘀𝗲𝘀𝘀𝗶𝗼𝗻𝘀.
that means the orchestrator keeps a continuous loop running (plan → delegate → handle ci/review feedback → continue), and you can still jump into any individual session to inspect and steer when needed.
@OpenAI symphony is the closest competition in orchestration-layer direction. in our view, the difference is not autonomy. it is that symphony’s orchestration layer is not truly agentic in the same way: it is not centered on an agentic manager layer continuously managing other agents end-to-end.
also important context: this is pure open source, not a composio product offering, and i mostly work on it on weekends.
~𝟱𝟬𝟬 ⭐ 𝗮𝘄𝗮𝘆 𝗳𝗿𝗼𝗺 𝟱𝗸.
once we cross 𝟱,𝟬𝟬𝟬 ⭐, core contributors get 𝗰𝗹𝗮𝘂𝗱𝗲 𝗰𝗼𝗱𝗲 𝗺𝗮𝘅 + 𝗰𝗼𝗱𝗲𝘅 𝗺𝗮𝘅 covered through anthropic/openai open-source grant programs (as per policy terms).
if you care about building this in the open, come join us.
github + discord in reply.
English
Retardmaxxing is the key to happiness
YET I CHOSE TO PLAY MATIKS 😭😭😭
English
@dishant0406 @aoagents @composio why don't you join us at 10 PM today and help take it forward?
discord.gg/M5fKq8pm
English
@agent_wrapper @aoagents @composio From internal tooling side project to 4.5K stars — the best open source projects start with scratching your own itch. Running parallel agents on real repos is a genuine pain point. Excited to see where this goes! 🔥
English
Will be talking about agentic orchestration and the future of knowledge work at MG Road, @neon_fund this Saturday at 11 AM IST
Thanks a lot for inviting, Shikhar Prateek!
looking forward to interacting with founders and builders in the AI infra space
luma link below
English
Italiano
does anyone have any tips on how to prompt/plan when trying to oneshot large projects, like 50K+ LOC?
English
English
English
@agent_wrapper The concept of agents rewriting their own orchestrator is a huge step for self-evolving systems. Curious to see the open source code!
English