You found openai.fm, typed a line in, and what came back sounded like a person instead of a GPS unit. Then you tried to use it for something real and hit the cap.
That isn't a flaw. It's the job. A demo exists to prove the tech works, so it hands you a taste and stops. Nothing saves, nothing is yours, and if the page comes down, which has already happened once, the whole thing evaporates.
uttrd is not a demo with a bigger allowance. It's a place to work.
Make it, then keep it
Every take stays in your recent list under the editor. Make three versions, play them back to back, download the one that lands.
Make a free account and they go somewhere better: a library that follows you across devices. Star a voice and it leads the lineup every time you come back. The take you made on your laptop at midnight is on your phone in the morning.
This is the part a demo structurally cannot do. It isn't a feature gap, it's a purpose gap.
Controls, not just a text box
18 voices, up to 4,000 characters per generation. Tap any voice to audition it before you commit.
When a read isn't landing, Tone, Energy, and Pacing are one-click levers. Or open the style box and describe what you want in plain English, which often gets you there faster. Hit Preview and you'll hear the opening before spending a full generation on the whole script.
If you want to go further on getting a natural read, we wrote up 7 script tricks that make AI voiceovers sound human.
What you walk away with
Download the MP3. No watermark. The audio is yours to use commercially, which covers monetized YouTube, Shorts, TikTok, podcasts, and client work.
There is a daily ceiling so nobody hammers the API into the ground, set high enough that normal use won't come near it. Not a handful of generations and done.
What's free, and what's coming
Every voice is free, permanently, no account required. That's not an intro price and it isn't going away.
All 18 voices are live now, and several of them take close direction on tone and character through the Tone, Energy, and Pacing controls. What's coming is length: scripts that run long and hold the same voice all the way through.
For the full picture of what uttrd does and doesn't do, see what is uttrd.
FAQ
- How do you say “uttrd”?
- “Uttered.” Same word, fewer vowels.
- How is uttrd different from openai.fm?
- openai.fm is OpenAI's demo page for its speech models. It caps you after a few generations and nothing you make is saved. uttrd is a tool for actually producing voiceovers: you download the MP3, keep your takes in a library, and shape the read with Tone, Energy, and Pacing controls.
- Is uttrd free?
- Yes. Every voice is free permanently, no account required. There is a daily ceiling to prevent abuse, set high enough that normal use won't reach it.
- Do I need an account?
- No. An account saves your takes to a library across devices, but you can generate and download without one.
- Can I use the audio commercially?
- Yes. Monetized videos, podcasts, client work, all fine.
- What file do I get?
- An MP3, downloaded straight to your device.