Use ScaledThought with LiteLLM

The Python SDK and the proxy both call /v1/chat/completions, so streaming, tools and response_format use the OpenAI request shapes.

Paste the prompt above into your coding agent, or run the command yourself in your project's directory. It creates the account, writes ST_API_KEY and ST_BASE_URL to ./.env (mode 0600) and adds .env to .gitignore. Only the US region is open; Canada is coming soon.

npx -y scaledthought@0.3 init --region us --source org-litellm-guide

Usage is prepaid, so a new account answers 402 insufficient_credits until its first top-up. This prints a payment link and opens it in your browser; nothing is charged until you pay on it. How prepaid credits work

npx -y scaledthought@0.3 billing topup --amount-usd 20
npx -y scaledthought@0.3 billing wait

npx -y scaledthought@0.3 models --live lists every model live in your region; the examples use deepseek-v4-1-flash.

import os, litellm

resp = litellm.completion(
    model="openai/deepseek-v4-1-flash",
    api_base="https://api.us.scaledthought.com/v1",
    api_key=os.environ["ST_API_KEY"],
    max_tokens=1024,
    messages=[{"role": "user", "content": "Reply with OK"}],
)
print(resp.choices[0].message.content)

Add the model to the proxy's config.yaml, then start it with litellm --config config.yaml. The proxy reads the key from its environment.

model_list:
  - model_name: deepseek-v4-1-flash
    litellm_params:
      model: openai/deepseek-v4-1-flash
      api_base: https://api.us.scaledthought.com/v1
      api_key: os.environ/ST_API_KEY

• Give the proxy its own sub-agent key with a daily budget, so one runaway client can't spend the account's balance: npx -y scaledthought@0.3 keys create --child --name litellm-proxy --budget-usd 5.

• LiteLLM: OpenAI-compatible endpoints, read 2026-10-10.
• ScaledThought: /llms.txt (the setup protocol) and the docs.