Text to Speech reads your text aloud with a natural-sounding AI voice and gives you the audio to download — for a voice-over, a video narration, an announcement, or simply to listen to something instead of reading it. It handles English, German, Persian and many other languages; the language is taken from the text itself, so there is nothing to select.
The text is sent to a cloud speech model, so this needs an account; each run uses 1 credit, which means the free plan's 30 monthly credits cover 30 runs. A run reads up to about 4,000 characters — roughly two pages — and anything beyond that is cut off, so split long scripts into parts.
How to use
Sign in and type or paste your text — up to about 4,000 characters per run.
Pick a voice: Alloy, Nova, Shimmer or Onyx.
Generate the speech, listen to it in the page, and download the audio file.
FAQ
Does it speak Persian and German?
Yes. The voice model detects the language of the text and reads it in that language — English, German, Persian and many more. Names and rare technical words are where mispronunciations are most likely, so listen through before you publish.
How long can the text be?
About 4,000 characters per run, which is roughly two pages. Anything longer is cut off at that point, so split a long script into sections and generate each one.
Can I change the speed, pitch or emotion?
Not in this tool — you choose one of four voices and the model sets the pacing. Punctuation helps: commas and full stops create natural pauses, and a new paragraph gives a longer one.
Why does it need an account and credits?
The speech is generated by a paid cloud model, not on your device. Each run uses 1 credit, and if generation fails the credit is refunded automatically.