Artificial intelligence news.
Straight from the original announcements.

Fish Audio's Drama 3 text-to-speech model arrives on Vercel AI Gateway

October 13, 2026By the AI News Bites Editorial TeamSource: Vercel
Fish Audio's Drama 3 text-to-speech model arrives on Vercel AI Gateway
Image: Vercel

Vercel has added Fish Audio's Drama 3 to AI Gateway. The text-to-speech model, currently in preview, lets developers direct voice performances with natural-language instructions written into the script.

Key points
  • Fish Audio's Drama 3 text-to-speech model is now available on Vercel AI Gateway, in preview.
  • Voice delivery is directed with natural-language instructions in square brackets and tags such as <whisper> and <emphasis>.
  • The model supports multi-speaker dialogue with a separate voice per character, and directions can be in a different language from the speech.
  • It is offered at 80% off for 30 days, through Nov 12.
  • Developers use the model name fish-audio/drama-3.

Directing speech from inside the script

According to Vercel, Drama 3 is aimed at voiceovers, character dialogue and audio storytelling. Instead of adjusting separate settings, users describe how a line should sound directly in the text.

Emotion and pacing can be described in square brackets. Sounds and pauses can be placed where they belong in the script. Selected words can also be wrapped in tags such as <whisper> and <emphasis> to change how they are delivered.

Vercel also notes that directions can be written in a different language from the spoken text. Drama 3 supports multi-speaker dialogue as well, with a separate voice for each character.

Getting started with the AI SDK

Developers call the model with the name fish-audio/drama-3. Vercel's example uses the AI SDK's experimental_generateSpeech function. The sample script passes the text [warm and reassuring] Take your time. <whisper>I am right here.</whisper>, a voice ID and an output format of mp3, then writes the returned audio to a file.

The model can also be tried in the model playground, and Vercel links to a page with model details.

Discounted pricing and what AI Gateway offers

Drama 3 is available at 80% off for 30 days, through Nov 12, according to Vercel. The source gives no regular price for the model.

AI Gateway is Vercel's unified API for calling models. Vercel says it covers tracking usage and cost, and configuring retries, failover and performance optimizations for higher-than-provider uptime. It also lists built-in custom reporting, Zero Data Retention support, budgets for API keys and routing rules.

Vercel states that AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests.

What it means for developers

For teams that already route model calls through AI Gateway, Drama 3 adds a speech option reachable with the same API and the same usage and cost tracking. Because the model is in preview, its behavior and availability may still change. The discount window runs only through Nov 12, so anyone planning to test it at the reduced rate has a limited period to do so.

Related news

Vercel Adds Parallel Decisions to AI Gateway for Free
For developersNovember 13, 2026

Vercel Adds Parallel Decisions to AI Gateway for Free

Vercel has made Parallel Decisions Fast from Parallel available through AI Gateway, with free access through November 23, 2026. The model searches the live web and…

Vercel Adds Kimi K3.1 Pro to Token Pass for Pro Teams
Products and toolsNovember 12, 2026

Vercel Adds Kimi K3.1 Pro to Token Pass for Pro Teams

Vercel has added Moonshot AI's Kimi K3.1 Pro to Token Pass, its Early Access subscription for Pro teams using coding agents through AI Gateway. The model supports…

Vercel adds 14 ElevenLabs audio models to AI Gateway, at half price for now
Products and toolsOctober 27, 2026

Vercel adds 14 ElevenLabs audio models to AI Gateway, at half price for now

Vercel has partnered with ElevenLabs to bring 14 ElevenLabs models to AI Gateway, covering transcription, speech synthesis, realtime voice agents, music, sound effects…