Scale Your Enterprise AI Infrastructure.
Across Every Model.
One AI infrastructure for every model your enterprise needs. Connect, route, and scale AI workloads with the flexibility, reliability, and control your business demands.
Free account · Prepaid model usage
Your AI, at a glance.
Every model. Every request. All in one place.
ONE OPENAI-COMPATIBLE API. A WORLD OF MODELS.
One API. More possibility.
Connect your existing OpenAI client. Access multiple models with one endpoint and one API key.
Keep your ideas moving.
Built-in provider fallback keeps requests moving when an eligible alternative is available.
Every rupiah, accounted for.
Track usage, set workspace limits, and manage prepaid credits in IDR. Top up through sales.
Build solo. Collaborate in scoped workspaces. Manage pricing and credits as a platform admin.
THE CONNECTION, IN MOTION
Scroll to exploreFollow the
possibilities.
- 01CONNECT
One key. A world of models.
Your request enters through one OpenAI-compatible endpoint. LIBRA connects it to the configured model providers.
- 02KEEP MOVING
A different route. The same API.
If a provider is unavailable, eligible fallback providers offer another path. Your integration stays the same.
- 03STAY IN CONTROL
Every connection, accounted for.
Track model usage and costs across your workspaces. Set limits and manage prepaid credits in IDR.
GET STARTED
Three steps to your first request.
From sign-up to your first API call, in one sitting. No new SDK, no rewrites.
- 01 / 03
Create an account
Sign up for free in minutes. Your organization starts with a Default workspace, ready for your team.
Create account - 02 / 03
Top up credits
Add prepaid credits in IDR. Usage is deducted per request, with limits you set for every workspace.
Contact sales - 03 / 03
Get your API key
Generate a key, point your OpenAI client to LIBRA, and start calling any model from one endpoint.
Sign in to console
from openai import OpenAI
client = OpenAI(base_url="https://<your-host>/v1", api_key="LIBRA_API_KEY")
client.chat.completions.create(model="libra/smart", messages=[...])The freedom to build with any model. The simplicity of one connection. We’re making AI more accessible with the visibility, flexibility, and control to move your ideas forward.
A foundation for MoMA tomorrow.
Our long-term vision is a Mixture of Models and Agents. Today, start with the model gateway: one connection, more possibilities.
FAQ
Questions, answered.
Everything you need to know before your first request. Still curious? Talk to our team.
What is LIBRA?
LIBRA is a model gateway: one OpenAI-compatible endpoint and one API key to reach multiple AI models. It adds provider fallback, workspace controls, guardrails, and prepaid credits in IDR on top.
Do I need to rewrite my code?
No. Keep your existing OpenAI client, change the base URL to your LIBRA endpoint, and use your LIBRA API key. Requests and responses follow the OpenAI format.
Which models can I use?
DeepSeek, Claude, GPT, Gemini, JiuTian, and ASIX-hosted models, plus aliases such as libra/smart, libra/fast, and libra/cheap. The model catalog lists everything currently available.
How does billing work?
LIBRA uses prepaid credits in IDR. Each request is deducted from your balance based on the model's token pricing, and you can see usage per key, per model, and per day. Top up through our sales team.
What happens if a provider is unavailable?
When a provider fails and an eligible alternative is configured, LIBRA routes the request through the fallback provider. Your integration and API key stay the same.
Can I control access for my team?
Yes. Organize keys into workspaces, assign roles (Owner, Superadmin, Admin, Member), set credit limits per workspace or key, and attach guardrails such as allowed models and max tokens.
Is it free to get started?
Creating an account is free. You only pay for model usage, drawn from your prepaid credit balance.
YOUR NEXT IDEA STARTS HERE