# ElevenLabs on AI Gateway

Source: [Vercel](https://vercel.com/blog/elevenlabs-on-ai-gateway)  
Feed7 permalink: https://feed7.dev/p/elevenlabs-on-ai-gateway-0z8kb7d  
Published: 2026-10-27T07:00:00.000Z  
Trust: Official Source (official_source)

## Why Included

Vercel AI Gateway adds 14 ElevenLabs models across transcription, realtime voice, music, effects, and dialogue under one API key, with shared cost and trace views. Realtime models use semantic VAD, not server VAD.

## Source Summary

AI Gateway adds **14 ElevenLabs models** across six audio types: transcription, streaming speech, realtime voice, music, sound effects, and multi-voice dialogue. Requests share one gateway key and common usage, cost, and trace views. Realtime voice models use one WebSocket session; the streaming pipeline can instead keep transcription, language, and speech models swappable.

## Practical Implication

Builders can choose between a modular speech pipeline and a single realtime model, and route audio alongside other models through the same gateway. Browser realtime setup requires a short-lived token, and the models support **semantic VAD** but not server VAD.

## Agent-Ready Context

AI Gateway adds **14 ElevenLabs models** across six audio types: transcription, streaming speech, realtime voice, music, sound effects, and multi-voice dialogue. Requests share one gateway key and common usage, cost, and trace views. Realtime voice models use one WebSocket session; the streaming pipeline can instead keep transcription, language, and speech models swappable.

Builders can choose between a modular speech pipeline and a single realtime model, and route audio alongside other models through the same gateway. Browser realtime setup requires a short-lived token, and the models support **semantic VAD** but not server VAD.

The listed 50% discount runs through **November 9**. The material does not provide comparative latency, quality, or cost results, so model selection still needs workload-specific evaluation.

## Connected Context

Feed7 judgment across 944 accumulated Signals:

This expands AI Gateway audio from individual speech routes to a broad ElevenLabs menu spanning transcription, realtime voice, music, effects, and dialogue. It also makes the architecture choice explicit: use one realtime session or keep transcription, language, and speech models swappable. The shared gateway provides common usage and trace views, but the source gives no comparative results for choosing a model.

- [Gemini 3.8 text-to-speech models now available on AI Gateway](https://feed7.dev/p/gemini-3-8-text-to-speech-models-now-available-on-ai-gateway-1ur27ch) — Both add speech-generation choices to AI Gateway, but Gemini divides TTS by generation style while ElevenLabs offers a wider audio surface and a realtime-versus-modular pipeline choice; neither source establishes a workload-independent winner.
- [Routing rules now available on AI Gateway](https://feed7.dev/p/ai-gateway-routing-rules-024ifzc) — Routing rules provide a gateway-level way to redirect or deny model choices, which could operationalize changes between the modular and realtime audio paths without changing application code.

## Context Map

- Layer: infra
- Domains: audio
- Topics: gateways, observability, model-selection

## Uncertainty

- The listed 50% discount runs through **November 9**. The material does not provide comparative latency, quality, or cost results, so model selection still needs workload-specific evaluation.

## Agent Instruction

Use this item as source-backed context. Do not invent claims beyond the linked source. If this item conflicts with another source, call out the conflict.