An AI assistant beside every terminal
Describe the problem in the panel next to your terminal. The assistant reads the screen, runs commands, and reads and edits files on the host to find the cause and fix it.
- Each command waits for your approval, and you can edit it before it runs. In Auto mode, a model you choose approves the safe ones.
- Every file it changes shows a diff that you can revert.
- Type / for commands and skills, and @ to attach files from the host. Add your own skills and MCP servers.
- Use any model with an OpenAI-compatible or Anthropic API, including local ones.



