Plurai is building the recursive self-improvement layer for AI agents. As agents take on real work across the enterprise, the ones that last will be those that learn from their own experience, growing more capable, more precise and more efficient with every run. Plurai unlocks that evolution for any company's agents, and lets every company own the intelligence behind them instead of renting it.
This is the 2nd launch from Plurai. View more
Plurai Simulation
Launching today
Find your agent’s worst bugs before your users do. One command in Claude Code, Codex or GitHub Copilot builds a digital twin of your agent’s world: synthetic personas and documents, plus mocked tools, APIs and databases, in sandboxes that never touch production.
Free Options
Launch Team / Built With
Render Sandboxes: a computer for every agentGive every agent its own machine. Starts in under a second
Promoted
Hey Product Hunt, Ilan from Plurai here.
Before Plurai, I built simulation for autonomous driving. Self-driving cars earned trust in millions of simulated miles before real roads. AI agents still skip that step.
Most agent testing puts a fake user in front of your agent and grades the chat. The transcript says "refund issued." Did the refund actually happen?
Plurai runs inside the coding agent you already use. Point it at your repo and your real, unchanged agent gets a digital twin: its tools, records and documents, rebuilt from your code, with one consistent world state. We mock what should be mocked and keep what matters live, like your RAG.
Personas that push back walk every flow, simple to brutal. Auto evals check what the agent did to the world, not what it said. Production failures replay from the exact step they broke, even after your agent changed. Every bug arrives with the conversation that proves it and a fix, re-run three times. Teams cut failure rates by 82%.
Connect your agent and start free: https://simulation.plurai.ai/connect.
Tell us what breaks first.
Hooray! This day has come 🎉 We were waiting for it so much! Meet Plurai Simulation 🚀
@maria_anosova good luck on the launch to you and your team
@ilankad23 An exciting launch!
@ilankad23 Go Team!
cool launch!! how close do the fake users get to real ones though, real customers do way dumber stuff than anything i'd script
@igorblink lol 100%, real users do the wildest stuff. that's why we don't script them. the personas get a goal and a personality and then just… wing it: typos, wrong booking codes, answering a question nobody asked, changing their mind halfway. plug in your prod traces and we learn how your users actually behave, so the sim goes from realistic to hyper-real. and when prod still finds something new, it flows back in: you can replay that exact convo as a test, and the new behavior enriches the personas so future scenarios include it too
@igorblink Thanks!
Hi everyone!
Use this to install on Claude code -
/plugin marketplace add plurai-ai/plurai-plugins
/plugin install plurai-simulation@plurai-plugins
/reload-plugins
/plurai-simulation:simulate
here for any questions! and here are the docs - https://docs.plurai.ai/
congrats on the launch ilan. what breaks first in my case is the order, the agent does all the right steps but in wrong sequence and the end state still looks fine. does the replay show a timeline of steps so i can see where it went off?
congrats on the launch ilan. for me the first thing that goes wrong is the order of steps. the agent does every right step but in a different sequence and the final result still looks ok. can the replay show the steps as a timeline so i can find where it started going wrong?
Congrats in the launch team!
Did you use your product to simulate on your product? 😅
@danshipit Haha! of course we did
and thank you!!!
Congrats on the launch! Testing what the agent actually does, not just what it says, is a smart approach. I'll try it out today.
@paulandrei Thanks!! cant wait to hear what you think about it