Free French Text to Speech
Free. No signup. 600 characters per clip.
What to expect from French synthesis
French output is generally fluent, and the interesting failures are in liaison, the linking of a normally silent final consonant to a following vowel. Whether a liaison is required, optional, or forbidden depends on grammar, so the model is inferring it, and it gets the common cases right while occasionally linking where a careful speaker would not. The accent is standard metropolitan French; no regional or Québécois option is available.
One thing that measurably helps
Punctuation carries more weight than you would expect. A comma before a subordinate clause reliably fixes a phrase the model otherwise runs together.
Other languages
Need the reverse? The free audio-to-text converter transcribes French audio, and French speech to text covers accuracy for that direction. Limits and paid plans are on pricing.
Need more than the free tool?
Wisprs transcribes full-length audio and video, with speaker labels, summaries, and exports to SRT, VTT, DOCX, and Markdown.
FAQ
Is this really free?
Yes, and there is no signup. Free previews are capped at 600 characters per clip with a daily limit per visitor, which is what keeps it free to offer. Longer text and premium voices need an account.
Which languages does it speak?
The voices are multilingual and take their language from the text you type, so most major languages work. We publish dedicated pages for Spanish, Japanese, French and Arabic because each one behaves differently enough to be worth explaining.
Can I pick a regional accent?
No. Output is one standard accent per language: neutral Latin American Spanish, metropolitan French, Modern Standard Arabic. If you need Castilian Spanish or a specific dialect, this is not the right tool.
Can I use the audio commercially?
The audio you generate is yours to use, including commercially. You are responsible for the text you submit, so do not paste copyrighted work you do not have rights to.
How long is the audio kept?
Download the file if you want to keep it. Free preview audio is not permanently stored and the link should be treated as temporary.
Why does it mispronounce names?
Names are the hardest case in every language, because the model is inferring a pronunciation from spelling with no context to go on. Spelling the name phonetically, or in kana for Japanese, usually fixes it.