Gemini 3.1 Flash TTS – with directed prompts

aanet 13 points 5 comments April 15, 2026
simonwillison.net · View on Hacker News

Discussion Highlights (3 comments)

aanet

The 3 examples, with three distinct styles, are fascinating. I'd like to see one with cockney accent, just for lulz

roscas

Hope they release an offline model for Ollama, a small one easy to work with for TTS in other languages.

Insensitivity

No matter what I wrote in the audio profile, AI Studio never followed it, regardless of scene or context. For example, I tried to get a male voice and kept getting female ones. Not sure if it's an AI Studio bug or I was doing something wrong.

Semantic search powered by Rivestack pgvector
4,686 stories · 44,186 chunks indexed