CloudBot gives you a fully configured AI employee with its own cloud computer in one click. Built on OpenClaw.
What you get: - Full Ubuntu desktop environment in the cloud - Pre-installed AI agent that sees the screen and controls the computer - 24/7 availability - your AI keeps working while you sleep - Uses your own API keys for AI models - Starting at $69/month
The AI can use VS Code, browse the web, run terminal commands, manage files - anything you'd do on a real desktop. I wake up to completed code reviews, finished research reports, and updated documentation.
Built this because setting up an AI agent with a proper working environment was always the painful part. Now it's just one click.
Demo video: https://www.youtube.com/watch?v=wSe4bvDMuKQ
Would love feedback from the HN community!
This is just a sign-up page for a waiting list.
A sign-up page cannot be a Show HN:
The product is fully functional - you can sign up and use it right away. We have a waitlist approval step because we need to manage cloud infrastructure capacity responsibly. Once approved (usually within hours), you get full access to spin up instances immediately.
Happy to fast-track if you want to try it out!
Interesting business model.
I have two questions:
1) Who do you see as primary competitors in this niche?
2) What is your model for managing liability in a situation where a naive, less technical (or plain lazy) customer lets one of these run a business without sufficiently verifying output? Your definition of due diligence may differ from theirs.
Great questions!
1) Competitors: There are players approaching this from different angles - some focus on code completion (Copilot), some on chat (ChatGPT), some on specific automation tools. What I wanted was an end-to-end solution: a cloud-based AI employee with its own persistent environment. That's the gap I'm filling.
2) Liability: Valid concern. The AI can make mistakes, and that's a real risk. We're implementing safeguards - you can watch tasks in real-time, pause/stop at any point, and we encourage users to review outputs before deploying to production. It's definitely not "set and forget" for critical work. Still iterating on making this safer.
Boy this sounds like the equivalent of Nightmare in Elm Street but for developers and instead of waking up to freddy you wake up to thousands of code reviews made by this thing.
Or a dropped database.
It looks like you're optimizing for meme value.
I can understand why you want an AI to use a desktop. But it's still absurd to use the least efficient interface possible for interactions with the outer world.
Having said that, fun project, good luck :) I am sure quite a few people would want to try it.
Thanks for the feedback and the kind words!
I'm curious - what would you imagine as a better interface? The desktop approach has overhead, but the advantage is it can use ANY software (VS Code, browsers, GUI apps) without needing custom integrations for each tool.
Would love to hear your thoughts on alternatives!
Feedback:
I’m not paying $70/mo to try this but there’s a price that I would that’s above free, maybe $5/mo and I’ll bring my claude/oai API key
If you had a one time “local” setup fee with deepseek/llama or whatever instead of the services, I’d probably pay more than $70
Appreciate the honest feedback! Almost all of the $69 goes to cloud infrastructure costs (the instance runs 24/7 with GPU access). I hear you on wanting a lower entry point.
Two things I'm considering: - Local version (bring your own hardware) - would eliminate the cloud cost - Cheaper tier with limited hours
For now, the value prop is "one-click cloud solution" but I'll definitely explore local deployment soon. Thanks for the input!
Why would you need a GPU for an AI managed instance? I guess it would useful for some workloads, but arguably not for most really.