Skip to content

Introduction

The ResetData AI platform provides OpenAI-compatible inference over a curated catalog of open models, with usage-based billing, running on Australian infrastructure.

  • Call models through an OpenAI-compatible API — chat completions, embeddings, reranking, image generation and audio transcription.
  • Choose models from a curated catalog of open-weight models, browsable at app.resetdata.ai/models.
  • Track usage and spend per request, against a prepaid workspace wallet.
  • Manage access with workspace roles, scoped API keys, and multi-factor authentication.

Serverless models are live and pay-per-token — pass the model’s ID and call it, with no provisioning. This is what the Quickstart uses.

Dedicated deployments run a model on GPU capacity reserved for you. This is currently arranged with our team rather than self-serve; model pages in the app include an enquiry option where dedicated serving is supported.

  • Get started — create an API key, make your first request, handle errors.
  • Models — the catalog, choosing a model, and the per-model parameters that affect how you call it.
  • Guides — task-focused how-tos for each endpoint.
  • Platform — rate limits and queueing, usage and costs, billing.
  • Your account — workspaces, roles, and account security.
  • Security — what we store, and what we don’t.
  • API reference — the complete, interactive endpoint reference.
  • Coming soon — features on the roadmap, and roughly how they will work.