Chat and completion routing
A single OpenAI-compatible interface routes requests to the model you specify, so switching models is a configuration change.
Model access platform
AI Token Hub
One OpenAI-compatible endpoint for leading model APIs, with unified keys, usage and billing under a single Singapore commercial relationship.
- Gateway
- OpenAI-compatible
- Billing
- USD
- Access
- Self-serve signup
Quick start
From key to first response in three steps
The gateway speaks the OpenAI Chat Completions protocol, so most existing SDKs work by changing the base URL and the key.
01
Create an account
Register yourself on the console and verify your email. No sales process is required before you start testing.
02
Issue an API key
Generate a key per project or per environment so usage stays attributable and keys can be rotated independently.
03
Point your client at the endpoint
Replace the base URL and the bearer token in your existing SDK or HTTP client. No other code changes are required.
curl https://maas.ubi-comp.com/v1/chat/completions \
-H "Authorization: Bearer $UBI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "<model-id>",
"messages": [
{ "role": "user", "content": "Hello" }
]
}'What you get
Commercial model access, not just an API key
Multi-model supply
Access to leading commercial and open-weight model families — flagship, mid-tier and small models — through a consistent commercial arrangement.
Single billing relationship
One consolidated USD invoice covering all model usage, instead of several foreign card-based subscriptions.
Pay-as-you-go
Usage-based billing only. You top up and spend against actual consumption, with no minimum commitment or subscription tier.
Model selection advice
Practical guidance on matching workload to model tier — where a small model is sufficient and where it is not.
Usage visibility
Per-project and per-team usage reporting so spend can be attributed rather than argued about.
Onboarding support
Key provisioning, endpoint configuration and migration help for teams moving from direct provider accounts.
Who it is for
AI product teams, software companies, agencies and enterprise innovation units that need reliable, billable model access from a regional supplier.
- Model coverage
- Multi-provider
- Billing currency
- USD
- Commercial model
- Pay-as-you-go
- Access
- Self-serve registration
Console
Built on a managed gateway
The platform runs on a unified AI API gateway with an administrative console, so operations that normally need custom tooling are available out of the box.
Image generation
Drawing endpoints are enabled alongside text models for teams that need both under the same key.
Asynchronous tasks
Long-running jobs are submitted and polled through the task interface rather than held on an open connection.
Real-time usage view
Consumption is visible in the console as calls happen, so spend can be checked at any time for internal chargeback and finance reconciliation.
Ecosystem
Works with the clients your team already uses
Because the endpoint is OpenAI-compatible, common desktop and web clients connect by entering the base URL and key. The console links several of them directly.
Any client, SDK or framework that supports a custom OpenAI-compatible base URL will also work.
- Cherry Studio
- Lobe Chat
- DeepChat
- AionUI
- OpenCat
- AI as Workspace
- CC Switch
- AQBot
Integration guide
Prerequisites, key creation, endpoint reference and worked request examples.
Read the docs ↗Tell us what you are building
Send us your workload profile and expected monthly volume. We will confirm the models available to you, current pricing and the fastest path to production.
Available models, rate limits and pricing are set by the underlying providers and change over time. Current rates are shown in the console before you spend.
Talk to us