Built StateSleuth — AI-assisted planning, coding proposals, human approval, and Playwright bug detection.
Built using Alan AI and open-model integrations. @tryAlanAI@prasad_pilla
🔗 GitHub: https://t.co/UAHE6txFlb
Today we're launching @tryAlanAI so developers can stop babysitting their coding agents.
We kept running into the same problem ourselves.
Even as agents got better at writing code, we were becoming glorified button-pushers for our own AI: starting every run, feeding them context, unblocking failures, and checking their work.
We built Alan to change that.
@PayToPin_com@prasad_pilla@tryAlanAI Absolutely! Safety matters, especially when AI agents can modify code. Better to handle rollback before things go wrong.
🚀 StateSleuth update for #Openfire!
Built an independent Playwright verifier that catches real bugs and captures screenshot evidence.
Next up: reliable AI-generated fixes! 🤖
@prasad_pilla@tryAlanAI#Openfire#BuildInPublic
@AvatarsMuse@prasad_pilla@tryAlanAI Rollback was definitely one of the trickier parts! Especially handling partial write failures and making sure the system reports accurately whether files were restored. Stale change detection was another key safeguard. Still more to test as I connect browser verification!
@navneetsuman_@prasad_pilla@tryAlanAI Great question! Planning to use Playwright for real browser interactions and assertions, with screenshots as supporting evidence.
@USASmarTekAI@prasad_pilla@tryAlanAI Yeah, that's exactly what I'm trying to build! Still figuring things out as I go, but the idea of having agents check each other's work is what makes this interesting to me.
@engineer_output@prasad_pilla@tryAlanAI Exactly! A successful build is only the first step. The real test is whether the feature behaves as expected. That's what I'm building StateSleuth to verify!
🚀 StateSleuth — building in public!
AI-powered coding verification: checking if features actually work, not just build.
Phase 2: Interactive demo with a planted bug 🐛
@prasad_pilla@tryAlanAI#Openfire#BuildInPublic
@mitansh_j07@prasad_pilla@tryAlanAI Exactly! A successful build doesn't guarantee a feature works as expected. StateSleuth aims to verify actual behavior with evidence. More coming soon!
Excited to join Openfire hackathon by Alan AI! 🚀
24 hours, solo building, and open-source models. Let's turn an idea into something useful and see what I can create! 🔥
@prasad_pilla@tryAlanAI#Openfire#Hackathon
Built Craftly, an AI-powered browser coding and sandbox IDE 🚀
I built it with React, Node.js, Docker, Kubernetes, AWS EKS, ECR, NGINX Ingress, and Mistral.
The AI agent can inspect a project, modify files, and generate working applications from natural-language prompts.
Update: CloudFront is finally enabled on my AWS account! 🚀
Huge thanks to @AWSSupport for helping resolve the account verification issue. I can now create my CloudFront distribution for Craftly.
One blocker down. 🔥
#AWS#CloudFront#BuildInPublic