OpenBot — governed “AI coworkers” that can operate a computer
LinkedIn post by Sandhya Ahuja
Open-source, self-hostable template for turning agents into governed “AI coworkers”.
OpenBot is an open-source, self-hostable template for turning agents into governed “AI coworkers” that can actually operate a computer.
How it works
- Each Bot gets its own container, Chromium profile,
/workspace, and logins — agents do not share files or browser sessions. - Browser actions, file operations, shell commands, MCP calls, and component use all go through a central gateway that:
- Resolves the target
- Evaluates a CEL policy
- Records an audit entry
- Only then executes — or refuses — the action
Framework-agnostic
Because a Bot is simply an AG-UI endpoint, agents built with frameworks such as:
- LangGraph
- CrewAI
- Mastra
- Pydantic AI
- Google ADK
- or custom code
…can be connected without making OpenBot dependent on one agent framework.
Human-in-the-loop
Humans can watch the Bot’s live screen, take control for things like login or 2FA, and hand control back afterward.
Links
- GitHub repo: https://lnkd.in/gEx8_Aue (via post)
Why this matters
The pattern: isolation per agent (container + profile + credentials) + central policy gateway (CEL) + full audit trail + human takeover for sensitive steps — addresses the two big enterprise blockers for computer-using agents: blast radius and accountability.