One integration. Many routes.
Use the OpenAI-compatible format and choose aion-auto or a catalog route.
Explore the feature COGNITIVE SWITCHBOARD
All your AI.
One intelligent route.
Connect models, build agents and run your AI with one API. From your first test to a product under your brand, with usage controls and billing in BRL.
OpenAI-compatibleBilling in BRLBring your own key
API GATEWAYAGENTSWORKFLOWSAPPS & STUDIOBYOKWHITE LABEL
ONE CONNECTED OPERATIONLess infrastructure to assemble.
Less infrastructure to assemble.
More room to build.
Integrate, create and track in one platform. Every step of your AI operation has a home.
Agents and automation
From conversation to execution.
Combine versioned agents, tools and workflows with limits and approvals.
Explore the featureApps and Studio
One place to create and organize.
Bring together projects, assets and multimodal jobs. Track each execution in your account.
Explore the featureControl and security
Your operation, under control.
Tenant isolation, permissions, budgets, logs and auditing for operational visibility.
Explore the featureROUTE CATALOG
Open full catalog ↗Choose how your AI works.
Routes from the public catalog. Aliases identify the execution path; availability and limits are checked in your account.
01Automatic
aion-auto
The right route for each task.
The classifier selects the execution class. Track usage in your dashboard.
View in catalog02Text
fox-fast
Quick answers, focused tasks.
Local Qwen route for classification, extraction and chat.
View in catalog03Text
fox-medium
More context for your conversation.
Local Gemma route for general tasks and assistance.
View in catalog04Code
fox-code
From the problem to the next commit.
Local Qwen Coder route for development and code analysis.
View in catalog05Text
fox-heavy
A route for more demanding tasks.
Local Qwen Heavy route. Check execution limits in your dashboard.
View in catalog06Advanced
aion-nvidia-nemotron-ultra
Reasoning on the NVIDIA Nemotron route.
Catalog fallback route. Availability and billing depend on your account.
View in catalogNo routes found. Try another search.
BUILT FOR DEVELOPERS
Your integration already knows the way.
Change the base URL, use your AionRouter key and keep the OpenAI-compatible format. Select aion-auto to route the task or choose a catalog route.
Read API guide ↗- Create your account and key
- Set the base URL and route
- Run and track usage
chat/completions
curl https://api.aionrouter.com/v1/chat/completions \
-H "Authorization: Bearer SUA_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "aion-auto",
"messages": [
{"role": "user", "content": "Olá, AionRouter!"}
]
}'PRICING IN BRL
Start with what you need.
Prepaid API with no monthly fee. Or connect your own keys with BYOK plans.
R$4,90
per 1 million tokensFast tasks, classification and extraction.
Prepaid API · no monthly feeGet started →R$14,90
per 1 million tokensChatbots, content and moderate analysis.
Prepaid API · no monthly feeGet started →R$59,00
per 1 million tokensReasoning, code and professional analysis.
Prepaid API · no monthly feeGet started →from R$299.00
per 1 million tokensPremium models and critical tasks.
Approval requiredGet started →R$49/month
Your own key for low volume.
Provider usage billed separatelyUnderstand BYOK →R$149/month
Your own key for production and higher volume.
Provider usage billed separatelyUnderstand BYOK →contact us
Custom limits and governance.
Provider usage billed separatelyUnderstand BYOK →Public pricing checked on October 10, 2026. Check prices, limits and conditions in your account before execution. View current pricing ↗
WHITE LABELThe infrastructure is ours.
The infrastructure is ours.
The brand is yours.
Configure visual identity, verified domains, plans, customers and access. Use the platform to build your own AI operation under your account activation conditions.
Explore White Label ↗ON-DEMAND CAPACITY
View current prices and terms ↗GPU for your next level.
Managed offer with billing in BRL, time controls and automatic shutdown. Capacity is subject to availability and confirmation.
RTX 4090 24GB
from R$8.90/h
BetaRTX 5090 32GB
from R$18.90/h
Subject to availabilityL40S 48GB
from R$22.90/h
Subject to availabilityH100 80GB
from R$48.90/h
Enterprise / reservationMinimum 1 hour · 15-minute blocks after the first hour · up to 3 prepaid hours · automatic shutdown.
FROM FIRST TEST TO OPERATION
Build. Run. Track.
- 01
Create your account
Access the platform and generate your API key.
- 02
Choose your path
Start with API, Assistant, an agent, workflow or Studio.
- 03
Run with control
Configure routes, budgets and permissions for your operation.
- 04
Track and improve
Review usage, costs, logs and results in the dashboard.
FREQUENTLY ASKED QUESTIONS
Before you begin.
What is AionRouter?
A platform for accessing AI routes through an API and operating Assistant, agents, workflows, Apps and Studio with usage and billing controls.
Do I need to rewrite my integration?
For OpenAI-compatible requests, use https://api.aionrouter.com/v1 as your base URL and an AionRouter key. Check supported features in the documentation.
How does aion-auto work?
The classifier selects an execution class and a compatible route. You can also choose a catalog route. Usage follows the class and account conditions.
How does billing work?
The API is prepaid in BRL with no monthly fee. BYOK has separate plans and provider usage is billed separately. Aion Heavy requires approval. Check current conditions before execution.
Can I use my own brand?
White Label operations let you configure identity, verified domains, plans, customers and access. Check documentation and your account activation conditions.
Are GPUs immediately available?
The managed offer depends on availability. RTX 4090 is in beta; RTX 5090 and L40S are subject to availability; H100 requires Enterprise reservation. Confirm the offer before booking.
COGNITIVE SWITCHBOARDYour next idea already has a route.
One API to get started. One platform to grow.