Quickstart

Hi Liberty speaks the OpenAI API format, so if you already have an OpenAI client set up, you're one base URL and one API key away from running open-source models instead.

1. Create an account and get a key

Sign up with a Southeast Asia-based account, and once it's reviewed you'll receive an API key by email. Keys are also viewable from your dashboard under Settings → API keys.

2. Point your client at Hi Liberty

from openai import OpenAI

client = OpenAI(
    base_url="https://api.hiliberty.ai/v1",
    api_key=HI_LIBERTY_API_KEY
)

completion = client.chat.completions.create(
    model="deepseek-v3.2",
    messages=[{"role": "user", "content": "Hello, Hi Liberty"}]
)

print(completion.choices[0].message.content)

3. Pick a model

Model names match what you see in the catalog — for example deepseek-v3.2, llama-4-scout, or qwen3-235b-a22b. Append -fast to any model name to use the Fast flavor instead of Base, e.g. deepseek-v3.2-fast.

Authentication

Every request is authenticated with a bearer API key tied to your account and billing.

• Keep keys server-side. Never ship an API key in client-side code or a mobile app bundle.
• Rotate on suspicion of leakage. Revoke and reissue a key instantly from the dashboard — old traffic on a revoked key fails immediately.
• Scope keys per project. Enterprise accounts can issue separate keys per project or environment for cleaner billing and audit trails.

Model flavors: Base vs Fast

Every model on Hi Liberty ships in two flavors. Both return identical output from the same checkpoint — the difference is purely serving configuration.

Switch flavors by changing the model name in your request — no redeploy, no separate endpoint to manage.

Rate limits

Starter accounts ship with generous defaults so you can build without hitting a wall during development.

Need higher limits sooner? Talk to sales →