One API for every frontier LLM

See the models Learn more

One OpenAI-compatible API for every frontier model — with automatic fallbacks, zero markup, and zero data retention.

Production Ready

Easiest, most reliable way to call multiple LLMs

Ship faster, spend less on tokens, and stop losing users to provider outages.

0% markup

Pay provider rates, not gateway rates. Competing gateways add 5% or more to every call.

Automatic fallbacks

Configure backup models per request. When a provider errors or stalls, your call still goes through.

Security by simplicity

Pay the exact same price as calling the model provider directly. No markup, no hidden fees, no minimum commitment. We make it simple.

OpenAI-compatible

Drop into any OpenAI SDK. Change a base URL and a model string, and everything keeps working.

Voice-native

Your LLM calls run where your transcription does. One less network hop on every turn.

Models worth using

Frontier models from OpenAI, Anthropic, Google, and open source. New ones added the day they launch.

Built for Voice AI

The LLM layer for Voice AI

One key, one bill for your entire Voice AI pipeline

Function calls and fallbacks fire mid-stream

Typed JSON out of the box, schema-enforced

Cached prompts cost less and respond faster

Live demo

Pick a model. Take action on your audio.

The same transcript, four different jobs, four different models — all routed through one endpoint.

Prompt

What is runner's knee?

claude-opus-4-7 412 ms · 47 tokens

Based on the transcript, runner's knee is a condition characterized by pain behind or around the kneecap. It is caused by overuse, muscle imbalance and inadequate stretching. Symptoms include pain under or around the kneecap and pain when walking.

Built deeper than the alternatives

Model

AssemblyAI LLM Gateway

OpenRouter

LLM Router

Bridges to

25+ models, every major provider

300+ models

OpenAI-compatible only

Speech-to-text integration

Native — no extra hop

—

Automatic fallbacks

Configurable per-model

—

Pricing

0% markup

5% markup

DIY infra

Same-day model releases

—

EU data residency

EU-resident, default-on for EU traffic

—

Zero data retention

Opt-in, per request

—

OpenAI SDK compatible

Playground

We’re not playing around, but you can

Put our Voice AI models to the test in our no-code playground.

Try it out

AI Speech-to-Text transcription in 98 languages

From Spanish to Korean, deliver accurate Voice AI in the languages your users speak.

🇪🇸 Spanish 🇵🇹 Portuguese 🇫🇷 French 🇩🇪 German 🇮🇳 Hindi 🇷🇺 Russian 🇳🇱 Dutch 🇯🇵 Japanese 🇮🇹 Italian 🇵🇱 Polish 🇺🇦 Ukrainian 🇮🇩 Indonesian 🇹🇷 Turkish 🇨🇳 Chinese 🇰🇷 Korean

Common questions

: AssemblyAI's LLM Gateway is a single, OpenAI-compatible API that routes requests to GPT, Claude, Gemini, Qwen, and other frontier models — with automatic fallbacks, no markup, and zero data retention available on every call.
: You pay the same price as the underlying provider — zero markup. No additional fees from AssemblyAI to use the Gateway.
: GPT (OpenAI), Claude (Anthropic), Gemini (Google), Qwen (Alibaba), Mistral, and 25+ other frontier and open-source models. New models are added the day they launch.
: Yes. Any OpenAI SDK or HTTP client works — change the base URL and API key, and the rest of your code is unchanged.
: Yes. EU customers can route requests through EU-hosted infrastructure for GDPR compliance. EU residency is on by default for traffic originating in the EU.
: Configure a primary model and any number of fallbacks. If the primary fails or rate-limits, the Gateway transparently retries on the next model in your list — no code changes required.
: Yes. Pass a header on each request, or configure it project-wide, to opt into zero data retention.