Gemini 3.8 Flash-Lite TTS: expressive text to speech, at volume
Gemini 3.8 Flash-Lite TTS is Google's workhorse voice model, built for the jobs measured in hundreds of files rather than one. It reads scripts across 101 languages, takes direction on how each line should land, drops in a laugh or a breath where you want one, and handles two speakers in a single pass. What it trades against the flagship is creative range, not expression. Available now in Picsart's AI Playground.
- Direct how each line is delivered
- Create dialogue with two speakers
Meet Gemini 3.8 Flash-Lite TTS, the one you run at scale
Most voice work is not one perfect read. It is a hundred product clips, a back catalogue that needs a second language, an assistant answering all day. Gemini 3.8 Flash-Lite TTS is Google's model for exactly that: high throughput, low latency and cost efficiency, built to take over production workloads from the previous generation. It is the volume tier of the Gemini 3.8 voice family, and it is the one you reach for when the question is how many rather than how good.
- Handle large voice production runs
Still gets direction, still laughs in the right place
The usual trade with a lighter voice model is that everything comes back level and lifeless. Not here. You still tell it how each line should land, still drop a laugh, a sigh or a pause exactly where the read needs one, and still stage two speakers in a single generation with each holding their own voice. It will even backchannel - the quiet "mm-hmm" that tells a caller someone is still listening, which is most of what separates a voice agent from a recording. Audio comes back as a standard WAV ready for an edit. The expression is intact - what you are giving up is the flagship's voice design, not its performance.
Dub a back catalogue without recording it twice
Google built this model for high-volume dubbing, and the maths is the point: 101 languages, a script at a time, at a cost that makes localizing an archive a decision rather than a project. Run a product library into the markets you have been putting off, give a course its second and third language, or push a week of social clips out in four. The same direction and the same vocal timing carry across every version, so the twentieth language sounds like the work, not like an afterthought.
- Dub a library into more languages
When to run Flash-Lite, and when to reach for the flagship
Run Flash-Lite when the job is measured in files: dubbing, read-aloud, voice agents, product clips, anything where the same script has to exist many times over. Reach for Gemini 3.8 Flash TTS when one performance has to carry - a designed character, an audiobook, a scene with acting in it. Inside Picsart sits alongside 150+ models, so you can voice a script here and switch models for the next one.
What you can make with Gemini 3.8 Flash-Lite TTS
Dubbing at volume
Take a library of finished video into new markets by generating the voice track per language rather than rebooking a studio for each one.
Gemini 3.8 Flash-Lite TTS FAQ
It is Google's high-volume text-to-speech model, released in September 2026 and built for throughput, low latency and cost efficiency. Google describes it as the workhorse of the Gemini 3.8 voice family, made to take over production workloads from the previous generation.
Generate AI visuals with GPT Image 2.5
Pro
Most popularAI tools for everyday creative work.
Ultra
Most powerfulHeavy AI usage for creators & teams.
Enterprise
Custom AI solutions for large organizations.
- Volume discounts on credit rate
- On-demand top-ups