gemini

Gemini Text-to-Speech

This sample generates speech with Gemini TTS and writes a playable WAV file.

Gemini TTS responses are returned as Genkit media parts. The media data may be raw PCM, usually:

audio/L16;codec=pcm;rate=24000

Raw PCM is not a WAV or MP3 file. The sample decodes the returned media data URI and wraps audio/L16 bytes with a WAV header before writing generated-tts.wav. This matches the JavaScript Gemini TTS samples, which decode media.url and convert the returned PCM bytes with a toWav helper.

Run the sample with a Gemini API key:

export GEMINI_API_KEY='<your-api-key>'
cd go
go run ./samples/text-to-speech/gemini

Invoke the text-to-speech-flow from Dev UI. The default model is googleai/gemini-2.5-flash-preview-tts, and the generated WAV path defaults to ./generated-tts.wav.

The dedicated *-tts models produce audio by default, so this sample configures speechConfig only. Use responseModalities: ["AUDIO"] with conversational Gemini models when you need to request audio as one of several possible output modalities.