Skip to Content
Introduction

HyperInfer API

One API for frontier models. Send requests in OpenAI Chat Completions, OpenAI Responses, or Anthropic Messages format — and use any model with any of them. The gateway translates between formats losslessly, streams natively, and bills usage against prepaid workspace credits with per-token pricing.

Base URL   https://api.hyperinfer.ai/api/v1
Auth       Authorization: Bearer <api key>
Formats    /chat/completions · /responses · /messages
Quickstart — first call in under 5 minutesAPI Reference — every endpoint, with a live playgroundPricing — models, credits, and fees

Why HyperInfer

  • Any format, any model. A Chat Completions request can call an Anthropic-slugged model; a Messages request can call an OpenAI-slugged one. Translation happens through a canonical internal representation — never string surgery.
  • OpenRouter-identical model slugs (anthropic/claude-sonnet-4.5, openai/gpt-4o-mini, …), so switching gateways is a base-URL change.
  • Streaming first. Native SSE per ingress format, usage reported on every request — streamed or not — including cached vs. uncached prompt tokens and reasoning tokens.
  • Server tools. Web search, web fetch, and PDF parsing run provider-side and are billed transparently as itemized ledger lines.
  • v1 is frozen. Everything under /api/v1 stays backward-compatible.

Versioning

All endpoints live under /api/v1. Breaking changes would require a new version prefix; none are planned. New capabilities are added under the same base path in a backward-compatible way.

Quickstart