A small bridge for your local models.

Your models, beyond localhost.

A small bridge between your frontend and the AI models running on your computer. Jobs wait in a persistent queue; your computer picks them up when it is online.

No open ports at home. Prompts and responses are encrypted between your browser and your local agent.

How it works

  1. Bring your computer. Install the agent alongside your model server.
  2. Make the connection. Pair it in the console, then add the SDK to your frontend.
  3. Send now. Read later. Follow the response as it arrives, or come back for the saved result.

For Ollama, LM Studio, llama.cpp, oMLX, vLLM and compatible text inference servers. You bring the models and the compute.

Get started

Currently a personal beta. Console access requires an owner key; public signup and paid plans are not open yet.