Vercel AI Gateway Adds 14 ElevenLabs Models

This turns AI Gateway into a more credible one-stop audio layer for developers, reducing the need to manage separate providers and billing paths.

• –Realtime speech-to-speech models support interruption handling, turn-taking, and WebSocket streaming for voice agents.
• –The catalog covers the full audio workflow, from transcription and synthesis to music and sound-effect generation.
• –Shared credentials, observability, routing, and spend controls simplify production operations across modalities.
• –The 50% promotion is useful for prototyping, but teams should benchmark costs after November 9 across character, minute, and session-hour billing.