AI Studio

Hyperstack AI Studio gives you two ways to run inference on AI models. Serverless inference runs models from a shared catalog of text and image models, through the playgrounds or an OpenAI-compatible API, and bills you only for what you use. Dedicated inference deploys an open-weight model on GPUs reserved for you, behind its own private endpoint. You can also ground answers in your own documents with knowledge bases, and connect AI Studio to the tools you already use through integrations.

Topics

Getting Started

Choose your path and start running models or building with the API in AI Studio.

Serverless Inference

Run catalog models on demand with no GPUs to manage, in the playgrounds or through the API.

Dedicated Inference

Deploy an open-weight model on GPUs reserved for you and get a private, API key-protected endpoint.

Models

The models available for inference, by modality, with pricing.

Knowledge Bases

Index your own documents so the Text Playground can answer from them and cite the passages it used.

Integrations

Connect AI Studio to popular developer tools, editors, and AI frameworks.

Billing

Usage pricing and cost transparency for AI Studio.