Hello Google Gemini team,
I am using Gemini TTS in Google AI Studio to convert long Moroccan Darija texts into speech.
Before purchasing paid credits, I would like to confirm whether the paid tier increases the maximum audio duration that can be generated from a single text input, or whether it only increases usage and rate limits.
Specifically, what is the maximum input text length and maximum audio duration for one Gemini TTS generation on the paid tier? Can several minutes of continuous Moroccan Darija narration be generated from one text input?
Upgrading to a paid plan only increases overall rate limits (RPM/RPD) and usage quotas, not the per-request token limit or continuous audio duration. See the Gemini Text-to-Speech (TTS) documentation for per-request model limits.
As per the TTS Limitations guide, voice quality and pacing can drift on generations lasting more than a few minutes.
The recommended best practice is: Split your Moroccan Darija text into smaller paragraphs/chunks, generate them sequentially, and Concatenate the resulting audio files together on the client side((e.g., using ffmpeg or pydub).