Generate speech quickly and at a lower rate with Google's Gemini 3.8 Flash-Lite TTS. Pick one of 70 preset voices, paste up to 5,000 characters, and create single-speaker audio for read-aloud features, drafts, and high-volume scripts.
Speed-first planning guide
These are planning previews, not stored model outputs. Use them to decide whether Flash-Lite fits the script before running the live generator above.
Visual voice planning preview
Here is today's summary. Three updates shipped this week, and one of them changes how you export your reports.
Visual voice planning preview
Your export is ready. Review the file, approve the final version, or send it back for revision.
Visual voice planning preview
Launch your next campaign in minutes. Write the script, choose a voice, and hear a draft before you commit.
When to use Flash-Lite
Flash-Lite is the speed- and cost-focused Gemini 3.8 TTS model on AISnapEdit. Flash is the fidelity-first option.
Rate
8 credits / 1,000
Charged per 1,000 characters, rounded up — a third less than Flash.
Voices
70 presets
The same use-case voice groups as Flash.
Languages
101 (per Kie)
Kie lists 101 languages for Flash-Lite; check yours before a long batch.
Output
WAV audio
Kie documents the output as WAV, 16-bit PCM, 24 kHz, mono.
Turn articles, help pages, and updates into audio without premium-rate spend.
Short, repeated messages that need one consistent voice.
Check wording and timing cheaply, then finish on Flash if needed.
Keep scripts compact, reuse one voice, and generate variants quickly.
Flash-Lite works best when the script is already concise and easy to say aloud.
Choose a voice from the group that matches the channel and keep it for every message.
Run short alternatives, compare clarity, and regenerate only the line that needs work.
AISnapEdit exposes two inputs for Gemini 3.8 Flash-Lite TTS.
Text
Every generation
Paste the final script, up to 5,000 characters. Credits are charged per 1,000 characters, rounded up.
Voice
Channel fit
Pick one of 70 preset voices. Kore is the default; reuse one voice so repeated audio sounds consistent.
Both Gemini 3.8 TTS models share the same 70 voices and inputs on AISnapEdit. Choose by what the audio is for.
Primary focus
Voice fidelity and expressive delivery.
Speed and cost efficiency.
Use Flash-Lite when you generate a lot of audio or iterate often.
Languages (per Kie)
130 languages.
101 languages.
Pick Flash if your language is outside Flash-Lite's list.
Rate on AISnapEdit
12 credits per 1,000 characters.
8 credits per 1,000 characters.
Flash-Lite keeps high-volume drafts cheaper.
Typical work
Narration, audiobooks, ads, and brand voiceovers.
Read-aloud features, bulk drafts, and short prompts.
Draft on Flash-Lite, finish on Flash if the delivery needs more range.
Each request reads the whole script in the selected voice. Dialogue between voices needs separate generations.
Generations use the 70 preset voices; there is no voice cloning or custom voice design on this page.
TTS cost is based on script length and shown before generation. If live pricing is unavailable, this page falls back to the recorded rate for this model.
credits per 1000 characters
Find answers to common questions about this model
It is Google's efficient Gemini 3.8 text-to-speech model for low-latency, high-volume speech. On AISnapEdit it turns a script into single-speaker audio in one of 70 preset voices.
8 credits per 1,000 characters, rounded up to the next 1,000. The generator shows the cost before you create audio.
Yes. Both models offer the same 70 preset voices, so you can draft on Flash-Lite and finish on Flash with the same voice.
Up to 5,000 characters per generation. Split longer material into sections and keep the same voice for each part.
Kie documents Gemini 3.8 TTS output as WAV audio (16-bit PCM, 24 kHz, mono), which you can play and download from your results.
Use Flash when the audio is a final narration, ad, or audiobook where expressive delivery matters, or when your language is not in Flash-Lite's list.
Paste your script, pick a voice, and check the credit estimate before you generate.
