OpenClaw is 500K lines of code, 53 config files, and 70+ dependencies. Good luck auditing that.
NanoClaw fixed that - 15 files, 3900 lines. But you still need Docker, a VPS, and a weekend to set it up.
We forked NanoClaw and replaced Docker with OnCell - each user gets their own isolated environment on NVMe with persistent storage, database, and search built in.
No server. No containers. One click.
Same WhatsApp/Telegram/Discord. Same persistent memory. But your data lives in a sandboxed cell that auto-pauses when idle and wakes in 200ms when you
message it.
Bring your Anthropic key → deploy on https://t.co/4yQhZjezDx → your AI assistant is running.
~$8/month. $0 when you're sleeping.
Open source: https://t.co/qP4XfPP6WK
Deploy: https://t.co/MIQIOJyatF
Just shipped one-click deploy for AI agents on OnCell.
Pick a template - Support Agent, Research Agent, or NanoClaw - enter your API keys, click deploy.
Your agent is live with a built-in playground in seconds.
Each agent gets its own persistent environment: storage, database, search. No Docker, no infra.
Try it: https://t.co/MIQIOJyatF
https://t.co/BOxCxlQ5d3
https://t.co/bVkcpYdZnk
https://t.co/qP4XfPP6WK
We open sourced a Perplexity clone.
Search the web, get streaming answers with sources. The entire backend is one file - no vector DB, no infra.
Built with Next.js + OnCell.
GitHub: https://t.co/t9JXsK2oYW
Demo: https://t.co/c94WpO3M3i
Just shipped: open source AI support agent built on OnCell
OnCell is infrastructure for AI agents - each customer gets their own isolated environment with storage, database, and search built in. No Docker, no Kubernetes, no vector DB to manage.
One API call and your agent is running.
To show what you can build with it, we open sourced a customer support agent. Upload your docs, get a chatbot that answers questions with citations. The entire backend is one file.
Demo: https://t.co/c7H1rlZVdh
GitHub: https://t.co/wfFcbNRkik
Site: https://t.co/CjOGn4p8s1
https://t.co/4yQhZjezDx is joining Canopy by @fdotinc.
5 weeks. 500 teams. SF + online.
We're building the infrastructure layer for AI agents - one API call gives each of your users their own isolated environment with files, database, compute, and a live preview URL.
If you're building agents, come find us.
Introducing https://t.co/4yQhZjezDx - ship AI agents without building infrastructure.
Every AI agent needs to give each user their own files, database, and compute. That's months of Docker, AWS, databases, isolation, and scaling - before
you write any agent logic.
Now it's one API call. Each user gets their own isolated environment. Your agent runs inside it with storage, database, search, and a live URL.
You write the agent. We handle everything else.
npm install @oncell/sdk
https://t.co/4yQhZjezDx
7/ We built this because we needed it.
We run Coding AI Agents for 100k users. Every user needs isolated compute for their AI coding agent. Building that infra took months. Now we're making it one API
call for everyone.