Connect

Use your private cloud from All AIs Desktop, from code, or bring your own GPUs.

All AIs Desktop

Sign in to All AIs Desktop with this same account. Open Use a paid model → All AIs private cloud: every deployment that is running here is listed, with its price. Pick one and the chat runs on it — no extra per-message charge, the hour is already paid for.

Desktop authenticates with your sign-in token, so there is no key to copy.

From code

Any OpenAI SDK works. Create a key under API keys.

from openai import OpenAI
client = OpenAI(base_url="/v1", api_key="aai-...")
client.chat.completions.create(model="<your deployment name>", messages=[...])

Full API docs

Bring your own compute ($0/hr)

Have a GPU box, a Mac Studio or an office server? Create a deployment with provider "Your own hardware", then run the connector next to your local OpenAI-compatible server (Ollama, llama.cpp, vLLM, LM Studio). It polls outbound only — no ports to open — and your team gets the same endpoint, key and usage log as a hosted deployment.

curl -o connector.js /api/connector
node connector.js --token <TOKEN from the deployment page> --target http://localhost:11434

Create a connector deployment