
Run agentic AI sessions with local MLX models, files, tools, and optional providers you choose.

Think
Local-first by design: run supported workflows locally and choose remote providers intentionally.

Prompts, files, and tool results can stay on your device for supported local workflows.

Work with documents, data, tools, code, and creative models from one controlled workspace.

Choose local MLX models where available, plus optional remote providers.
Think
Choose local models and provider boundaries

Use open models for local workflows that are customizable, inspectable, and optimized for your device.
For supported local workflows, prompts, files, and tool results stay on-device instead of going to hosted AI services.
Use remote providers intentionally when a workflow needs them.
Think
Send product feedback, unclear copy, or workflow gaps. Join Discord or email the team.