New: Claude Opus 5.5 & Sonnet 5.5, GPT-6 Sol & Luna, Grok 4.7 on Onysoft

The AI API Gateway

Access 749 AI models through a single API. OpenAI-compatible endpoint, transparent pricing, and corporate invoicing.

Get Started View Pricing

NEW Onysoft AI Gateway Node

AI on your own computer

Run open-source language models on your own graphics card — or on your CPU on Windows if the GPU isn't enough — and use them for free from the dashboard chat or with your personal API key, on the same endpoint as every other model. Other users' requests never reach your device, and your computer opens no inbound ports. Once your GPU Sharing Program application is approved, the app is available in your dashboard.

Windows 10/11 · macOS (Apple Silicon)
Gemma 4, Qwen3, Mistral, Phi-4 and other Apache 2.0 / MIT licensed models
Free to use on your own device

GPU Sharing Program How does it work? Read the announcement

Discover the Power of Onysoft AI Gateway

Try AI right now — no sign-up required info This demo runs on Onysoft AI Gateway. No sign-up required; a daily trial limit applies.

Hi! 👋 I'm the Onysoft AI demo assistant. How can I help you?

AI Models

749

Model Makers

Billed in TRY

Support

TR / EN

⚡

OnyRouter — Automatic Model Selection Per Request

NEW

Stop deciding which model to use: OnyRouter analyzes every request, classifies the question type and routes it to the best-fit model. Routing is free; you pay only the selected model's price. model: "onysoft/auto"

How It Works Documentation

hub

All Models, One API

Access 749 AI models through a single OpenAI-compatible endpoint. Use any provider without changing your SDK.

speed

High Availability

Uninterrupted service with a 99.9% uptime guarantee. Automatic load balancing and failover support.

payments

Transparent Pricing

No hidden fees. Pay in your local currency at live exchange rates. Pay as you go and get a corporate invoice.

shield

Data Security

TLS encryption in transit and data processing under KVKK (the Turkish data protection law); you can revoke and regenerate your API keys from the dashboard.

Why Onysoft?

One OpenAI-compatible API with billing in Turkish lira: an AI gateway built for developers.

code

OpenAI-Compatible

Access every model without changing your existing OpenAI SDK.

receipt_long

Corporate Invoicing

Tax-compliant local e-invoicing for your company.

support_agent

24/7 Support

Get instant support whenever you run into technical issues.

Popular Models

Instant access to the most advanced models.

View All arrow_forward

smart_toy

OpenAI: GPT-6 Astra

1.1M context

₺0,74

per 1K tokens

psychology

Anthropic: Claude Fable 5.1

1.0M context

₺0,74

per 1K tokens

psychology

Anthropic: Claude Opus 5.5

1.0M context

₺0,30

per 1K tokens

auto_awesome

Google: Gemini 3.8 Flash

1.0M context

₺0,06

per 1K tokens

psychology

Anthropic: Claude Sonnet 5.5

1.0M context

₺0,15

per 1K tokens

rocket_launch

SpaceXAI: Grok 4.7

500K context

₺0,15

per 1K tokens

Supported Providers

Models from the world's leading AI companies, all under one roof.

OpenAI

Anthropic

Google

Frequently Asked Questions

What is Onysoft AI Gateway?

Onysoft AI Gateway is an AI API platform that gives access to AI models from different makers through a single OpenAI-compatible API and a single key, with billing in Turkish lira. Text, image, video, audio and embedding models are used from the same account.

How does pricing work?

Prices are per model: per 1 million input and output tokens for chat models, and per unit (image, second, generation or 1,000 characters) for image, video and audio models, all listed on the /models page. Usage is deducted from your balance in TRY at the central bank (TCMB) rate at the time of the request; there is no subscription or monthly commitment.

Will my existing OpenAI code work with Onysoft?

In streaming mode, yes: set base_url to https://api.onysoft.com/v1, use your key with the sk-ony- prefix, and put the model ID from the catalog in the model field; streaming responses arrive in OpenAI chunk format. Non-streaming responses come back in a success/data envelope, with the OpenAI-format response object in the data field.

Which models can I access?

You can access the Anthropic Claude, OpenAI GPT, Google Gemini, xAI Grok, DeepSeek, Qwen, Meta Llama and Mistral families, along with image, video, audio and embedding models. The current list and prices are on the /models page.

Are there free models?

Yes. Nothing is deducted from your balance for models whose ID ends in :free or :tr; a daily request limit applies to them. For production work that has to run without interruption, paid models are recommended.

How do I get started and how do I pay?

Create an account, generate an API key in the dashboard and add balance; usage is deducted as you go. You do not need separate contracts with the model makers.

How is my data protected?

All traffic is encrypted with TLS in transit. You can revoke and regenerate your API keys from the dashboard. Details on the processing of personal data are in the Privacy Policy.

Start Building Today

Create your free account and start using it right away.

Create Your Account