Overview
What THROUGHPUTS is, what it does, and why you'd build on it.
One API key. Every major model. Transparent per-token pricing.
THROUGHPUTS is a unified API gateway that routes requests to every major AI provider — OpenAI, Anthropic, Google, Meta, Mistral, Cohere, DeepSeek, Qwen, and more — behind a single OpenAI-compatible endpoint. You integrate once, switch models in one line of code, and pay per token at bulk-negotiated rates.
What you get
- OpenAI-compatible API — point your existing OpenAI SDK at
https://api.throughputs.dev/v1and it just works. - Every major model — chat, embeddings, image, audio, multimodal. New models land within hours of release.
- Transparent pricing — per-token, no markup. The same price the provider charges, passed straight through.
- Routing & failover — automatically retry on a cheaper qualified model if your primary is down or slow.
- Startup program — 30% off for 12 months for qualifying startups.
- Reseller / white-label — deploy the whole marketplace under your own brand.
Where to go next
Quickstart
Send your first request in under five minutes.
API Reference
Endpoints, parameters, and response shapes.
SDKs
Official clients for Python, TypeScript, and Go.
Models & Pricing
How per-token billing works and what you'll pay.
Design principles
- Prices are the hero. Every catalog page leads with per-token cost, not marketing copy.
- One credential, every model. No per-vendor contracts, no procurement delays.
- Failover by default. The routing layer picks the cheapest qualified model if your primary is unavailable.
- Honest metering. You see the same token counts the upstream provider reports — nothing invented, nothing hidden.