oncell.ai

Developers building AI agents hit the same wall: every user needs their own files, database, and compute. That means Docker, S3, isolation, scaling months of infra before any agent logic. oncell makes it one API call. Each user gets their own isolated environment with file storage, database, shell access, search, a live preview URL, and built-in observability. Your agent runs inside it - direct local access, no network hops.

Intercom — Startups get 90% off Intercom + 1 year of Fin AI Agent free

Startups get 90% off Intercom + 1 year of Fin AI Agent free

Hey Product Hunt! 👋 I'm Anup. I built oncell because I needed it. I run an AI coding agent with 100k users. Each user needs their own isolated environment with files, database, and compute. Building that infrastructure took months. ECS, EFS, DynamoDB, Step Functions, security, scaling - all before writing any agent logic. I talked to other founders building AI agents. Everyone was solving the same problem. So I extracted our infrastructure into a platform. oncell gives each of your users their own isolated environment. Your agent code runs inside it. One API call. No Docker, no AWS config, no database management. What makes it different: - Agent code runs INSIDE the environment (not calling remote APIs for every file read) - Secrets injected as env vars (never written to disk) - Idle environments pause automatically, resume in 200ms - Built-in observability - every agent step logged - Live preview URL per environment The SDK is on npm (@oncell/sdk), docs are at oncell.ai/docs, and there's a demo agent that generates web pages from natural language. Would love your feedback. What would you build with per-user AI environments?

Hey everyone! I'm Anup, the maker of oncell.ai. Wanted to share the story behind it.

Why I built this