AI Voice Generator
Turn text into speech: paste your script, pick a voice and download a natural voiceover as an MP3 a few seconds later, for a YouTube video, a course, an ad or a podcast intro. Six models read English, including ElevenLabs, Gemini TTS and MiniMax, with 28 female and male voices to choose from. You can also clone your own voice from a 10 to 30 second sample. A recording costs 3 to 9 credits, so the 10 free credits you get at sign-up cover your first voiceovers.
What you get in the AI voice generator
Six text-to-speech models
ElevenLabs v3 and Multilingual v2, Gemini TTS, MiniMax Speech 2.8 HD, Chatterbox and Kokoro. On the English site, Fattly tells the model your text is English instead of letting it guess.
28 voices, female and male
14 Gemini voices, 7 from ElevenLabs and 7 from MiniMax, from calm to expressive. ElevenLabs v3 understands emotion tags, and you can play the ElevenLabs voices before you generate.
Your voice from a short sample
Upload 10 to 30 seconds of your speech and the model reads your text in your voice. A saved voice costs less after that: 40 credits per recording, and it also works in video ads and dubbing. A consent declaration is required.
MP3 ready for your editor
Download an MP3 for a video, course, podcast or ad. Commercial use within the Terms, a subscription is optional, and credits never expire.
How to make an AI voiceover in 4 steps
- 01
Paste your text
Up to 1,000 characters per recording, roughly a minute of speech (300 in Chatterbox). Write numbers and abbreviations out in words, so the voice reads them the way you mean.
- 02
Pick a model and a voice
Six models read English, each with its price shown: 3 to 9 credits. You can play the ElevenLabs voices before you generate.
- 03
Generate and listen
The recording is usually ready in a few seconds. Something sounds off? Adjust the punctuation or how a number is written and generate again.
- 04
Download the MP3
The file also stays in your account history for 30 days. If a generation fails for technical reasons, your credits come back.
What an AI voiceover is good for
Narration for videos and courses, podcast intros, ads and, with your saved voice, dubbing for short clips.
How much an AI voiceover costs
| Kokoro, English only, up to 1,000 characters | 3 credits |
|---|---|
| Chatterbox Multilingual, up to 300 characters | 4 credits |
| Gemini TTS, up to 1,000 characters | 6 credits |
| MiniMax Speech 2.8 HD, up to 1,000 characters | 9 credits |
| ElevenLabs v3 or Multilingual v2, up to 1,000 characters | 9 credits |
| Recording in your voice from a sample (cloning) | 150 credits |
| Each next recording in your saved voice | 40 credits |
| Dubbing a clip of up to 10 s in your saved voice | 35 credits |
| “Hands + voiceover” video ad in your saved voice | 190 credits |
| Background music: 15, 30 or 60 s | 12 / 20 / 35 credits |
The 10 free credits you get at sign-up (no card) cover three Kokoro recordings, one ElevenLabs v3 recording, or two shorter tests: Gemini TTS and Chatterbox. You pay per recording, not per character, so a short sentence costs the same as 1,000 characters. A subscription is optional, credits never expire, and a generation that fails for technical reasons costs nothing.
All packs and plansAI voice generator FAQ
Is the AI voice generator free?
To start, yes, within limits. You get 10 credits at sign-up with no card, and a recording costs 3 to 9 credits, so your first voiceovers are free. More recordings and voice cloning (150 credits) need a credit pack; the smallest has 150 credits, for example 25 recordings on Gemini TTS. There's no unlimited free plan.
Can I use it online without signing up?
It runs in your browser, with nothing to install, but generating requires an account with a confirmed email address. Sign-up is free and doesn't need a card.
Which voice should I pick for English text to speech?
For a female voice, start with Gemini TTS (14 voices, 6 credits). For a male voice, try MiniMax Speech 2.8 HD (Casual Guy or Deep Voice Man) or ElevenLabs (George, Brian or Charlie), 9 credits each. ElevenLabs v3 gives the most emotion, and Kokoro (3 credits) is the cheapest for drafts. The surest way to choose is to generate the same short paragraph in two models.
How much text can it read at once?
Up to 1,000 characters per recording, roughly a minute of speech, or 300 characters in Chatterbox. Split a longer script, such as the narration for a video or a course lesson, into parts and join the MP3 files in your editing app. Each part is a separate recording.
How do I clone my own voice?
In the My voice tab, upload 10 to 30 seconds of clear speech (MP3, WAV, OGG, M4A or AAC), type your text and tick the declaration that it's your voice or that you have the speaker's consent. The first recording costs 150 credits and each next one in the saved voice 40. You can also use the saved voice in an Ad Studio ad (+40 credits on top of the video) and to dub a clip of up to 10 seconds (35 credits). A saved voice you don't use for about a week can expire.
Can I generate a celebrity or cartoon character voice?
No. Fattly has no celebrity or famous character voices, and our Terms prohibit impersonating other people. You can clone only your own voice or the voice of someone who has expressly agreed. For original characters, pick a fitting voice from the list.
Which languages does it speak?
English, Polish, German, Spanish and French. The voice language follows the site language, so to get a voiceover in Spanish, for example, switch the site to Spanish first. Kokoro is the exception: it speaks English only.
Can I use the voiceovers commercially?
Yes, within our Terms and the model providers' terms: in ads, monetized YouTube videos, courses or client work. Fattly claims no rights to the recordings you generate, and you are responsible for what the voiceover says.
Guides
Give your words a voice
Make your first AI voiceover in seconds: 10 free credits, no card required.