Get Started With Cloudless AI

Send requests to models running on your machines through one OpenAI- and Anthropic-compatible API.

An application sends an OpenAI- or Anthropic-compatible request through the Cloudless AI Router to an eligible node agent inside the enterprise network. The model runs locally, and the response returns through the router. Administrators manage routing and keys in the dashboard; cloud fallback is optional per routing rule.

Get started

1. Create an account

Create an account in the Cloudless AI dashboard.

2. Install your node

Install Node.js 20 or later and npm, then install and start the node agent on a macOS, Linux, or Windows machine:

npm install -g @cloudless-ai/node-agent
cloudless-node start

Follow the browser sign-in prompt. When it finishes, confirm the machine appears online in Nodes in the dashboard.

Nodes dashboard showing a demo workstation online with the llama3.2 model loaded.

3. Configure a local model

In the dashboard, open Nodes, select your machine, and install a model. Then open Routing and create a rule that maps the model name your application will request (for example, my-model-alias) to the installed node model.

Routing page showing a rule that maps gpt-* requests to the local llama3.2 model, with gpt-4o-mini selected for cloud fallback.

4. Create an API key

In Settings > API keys, create an Inference key. Copy it when it is created and store it as a secret; the dashboard will not show the full key again.

API Keys page showing a demo inference key with a masked preview.

Set it in your shell before making a request:

export CLOUDLESS_API_KEY="<your-inference-key>"

5. Make your first request

Replace my-model-alias with the model name from your routing rule. The x-cloudless-node response header identifies the node that served the request.

curl
Python

curl -i https://api.cloudless-ai.app/v1/chat/completions \
  -H "Authorization: Bearer $CLOUDLESS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "my-model-alias",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Install the OpenAI SDK with:

pip install openai

Then run:

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.cloudless-ai.app/v1",
    api_key=os.environ["CLOUDLESS_API_KEY"],
)

response = client.chat.completions.create(
    model="my-model-alias",
    messages=[{"role": "user", "content": "Hello!"}],
)
print(response.choices[0].message.content)

Get started​

1. Create an account​

2. Install your node​

3. Configure a local model​

4. Create an API key​

5. Make your first request​