OpenBot

OpenBot turns AI models into a team of persistent agents that work together on your computer. Instead of one chatbot answering questions, you get named agents with individual instructions, workspaces, and conversation histories. They message each other, hand off tasks, and review each other's work, so a bug found by one agent can be fixed by another and merged by a third, without you orchestrating every step.

The app is free for personal use and runs on macOS, Windows, and Linux. Workspaces, conversations, attachments, and browser data live on your machine, though agents still connect to the model providers you choose.

What agents can do:

• Use plans you already pay for. Connect Codex, Claude Code, Gemini, Grok, OpenCode, Cline, Cursor, or any OpenAI-compatible endpoint, including local models via Ollama or LM Studio.
• Built-in browser. Agents open, read, and control web pages inside a browser that's part of the app, so they can interact with live sites without extra setup.
• Task queue. Send more work while an agent is busy. Messages wait in a queue you can pause, resume, or cancel.
• Automatic failover. When a plan hits its usage limit or a provider goes down, an agent switches to a backup model and keeps its chat and workspace intact.
• Agent persistence. An agent keeps its full context when you restart the app or switch it to a different provider.

Each agent has its own name and instruction set. You describe what you want in a prompt, review the generated instructions, and the agent is ready. Teams grow naturally: add more agents and they coordinate around whatever work is in flight.

For teams that need always-on availability, a managed server option runs OpenBot outside any single laptop, with shared or per-employee environments and custom integrations into existing tools and APIs. The source code is available under the PolyForm Noncommercial license, so you can audit what the app does, but commercial use requires a paid license.