prateek
@agent_wrapper
Followers
4K
Following
6K
Media
266
Statuses
5K
Orchestrator of agent orchestrators (@aoagents) @Composio
SF
Joined April 2019
I built an AI system that builds itself. Not metaphorically. The AI agents literally rewrote their own orchestrator, tested it, reviewed their own code, fixed their own CI failures, and shipped it. Introducing Agent Orchestrator. Open source. Link below.
53
169
475
@suraj_markup @aoagents we used Agent Orchestrator to build Agent Orchestrator. on a daily basis we run agentic workflows that self-correct, self-heal, and self-improve. if this kind of work excites you, we're hiring across SF and Bangalore. ⭐ https://t.co/ChYGa0qivo 📖 deep-dive:
github.com
Agentic orchestrator for parallel coding agents — plans tasks, spawns agents, and autonomously handles CI fixes, merge conflicts, and code reviews. - ComposioHQ/agent-orchestrator
0
0
1
went for a walk today. got a push notification on my phone: "session needs review. ao-52: idle." @suraj_markup built a mobile app for Agent Orchestrator(@aoagents). i can see all my agents running, their PRs, CI status. right from my phone. 8 agents running right now. 3 waiting
1
1
6
Hey everyone, just a quick update: -> Two people showed interest in this app yesterday, which is great! -> Spent my whole night working on polishing the app and making it ready for production. -> Finally it was ready by 6:00 AM. I've already given it to those guys. -> The PR
1
2
7
@aoagents we used agent orchestrator to build agent orchestrator. on a daily basis we run agentic workflows that self-correct, self-heal, and self-improve. if this kind of work excites you, we're hiring across SF and Bangalore: https://t.co/hWstcYCr02 repo: https://t.co/ChYGa0qivo full
github.com
Agentic orchestrator for parallel coding agents — plans tasks, spawns agents, and autonomously handles CI fixes, merge conflicts, and code reviews. - ComposioHQ/agent-orchestrator
2
0
9
agent orchestrator(@aoagents) lets one person manage 30+ parallel coding agents.⭐2,700+ github stars and 1M+ impressions since we open sourced it 8 days ago. but i want to talk about what made it possible. it started as 2,500 lines of bash managing tmux sessions and git clones.
7
8
147
recursive self improvement has taken off, there is no going back now
1
0
7
AO used itself to ship v0.2.0. the dashboard was loading in 51 seconds. an agent found the root cause (18 concurrent ps calls competing for resources on a 1000+ process machine), cached it, and brought it down to 6 seconds. another agent made the dashboard update in real time
𝘃𝟬.𝟮.𝟬 — AO improved itself and shipped a new release. 4 PRs. All written by parallel agents. Reviewed by humans. Merged by the orchestrator. 𝗪𝗵𝗮𝘁 𝗰𝗵𝗮𝗻𝗴𝗲𝗱: 𝟵𝘅 𝗳𝗮𝘀𝘁𝗲𝗿 𝗱𝗮𝘀𝗵𝗯𝗼𝗮𝗿𝗱 51s → 6s. Root cause: 18 parallel ps calls competing for resources.
4
1
13
full article: https://t.co/5SUoZvZtt6 examples in the post: • parsed hundreds of credit card transactions across 3 cards in 20 min • built an automated growth engine that surfaces 50+ tweet opportunities per day • 4 blog posts in 2 weeks, each with inline SVG diagrams •
0
0
2
knowledge work used to mean doing the work. research, execution, implementation. now it means knowing what work to do and whether it was done right. i run 30 parallel coding agents. haven't written production code in weeks. but i work the same hours. the work transformed, it
5
0
12
what a read, literally sent chills down my spine.. great job @suraj_markup what a way to leverage Agent Orchestrator.
The Agent Orchestrator removed the engineering bottleneck. The mobile app removed the desk. Multiple AI agents run on my machine. Every stuck session, every merge approval, every review comment used to pull me back to my laptop. Now my phone buzzes. I handle it from wherever I
0
2
21
Funny how just a week ago I was astonished by how good the models and agents have become, but today again I feel like they are just too damn slow.
2
0
4
wow this is amazing, thanks a lot for contributing @suraj_markup !
-> 3 agents running in parallel. 39 commits. 183 tests. 16 issues closed. 1 day. -> Claude built Codex support. Codex made itself better than Claude. -> My first open-source contribution to Agent Orchestrator.
0
0
6
@nikitabier @arshdeepsinghh @VastoLorde95 @avaitopiper @pranjalssh guys I'm going for it. Hoping X's bot detection is not broken. @elonmusk 🫡
0
0
1
0
0
0
2
0
1
hi @nikitabier! wanted to know your thoughts on smth In this age it quite trivial to setup agentic workflows that find you tweets of ideal customers primed for a sale if an individual were to generate both this list and the responses agentically, (100s) and then manually copy
2
0
5
so something like openclaw but packaged better?
What has Perplexity been up to last two months? We've silently been working on the next big thing: Perplexity Computer. Computer unifies every current capability of AI into a single system. Files, tools, memory, and models, orchestrated together, working for you.
2
0
6