Hey Product Hunt 👋
I'm Ruslan, solo maker on this.
I kept hitting the same wall with AI agents: they only work while you're watching. Close the tab and the task dies. Shut the laptop and the schedule never fires. Which rules out basically everything interesting, because the useful tasks are the ones that take longer than I'm willing to sit and watch.
So Krowoc deploys a micro-VM for the agent instead of running it inside your session. Real shell, filesystem, headless Chrome. Start a task from your phone, put the phone in your pocket, come back when it's done.
Each user gets their own isolated VM. That's mostly a security decision — an agent holding shell access plus your Gmail and Notion tokens is the highest-trust software you'll run, and I didn't want one prompt injection off a browsed page reaching anyone else's keys.
Models are bring-your-own-key: Anthropic, OpenAI, open weights through Fireworks, OpenRouter, or local Ollama. No proxy in the middle, nothing marked up.
Apache 2.0 and self-hostable — github.com/nikonr7/krowoc
Still rough in places: no Bedrock/Vertex auth yet, scheduled triggers are the newest piece, and I'm the only one on it.
The thing I actually want to know from you: what's the first task you'd hand to an agent and genuinely walk away from? Whatever people answer is what I'm building the next three months around.
Report
No reviews yetBe the first to leave a review for Krowoc