SpeakScribe now clones voices. Record 10–60 seconds of your own voice, and Voice Clone builds an AI copy of it that can read any script aloud — in more than 80 languages. It is available today on the Pro and Agency plans.
What Voice Clone does
Stock text-to-speech voices are great when you need a neutral narrator. But if your audience knows your voice — from your YouTube channel, podcast or course — switching to a stranger's voice breaks the connection. Voice Clone keeps it: you write the script, and it comes back in your own voice.
- Sample: 10 seconds minimum, 30–60 seconds recommended, from one speaker.
- Input: record in your browser, or upload MP3, WAV, M4A, OGG, WEBM, FLAC or AAC.
- Ready in: usually 1–2 minutes.
- Languages: 80+, including generating speech in a language you didn't record in.
- Output: MP3 at 128 kbps, up to 5,000 characters per render.
How to clone your voice in three steps
- Record or upload a sample. Read a paragraph for 30–60 seconds in a quiet room. A phone or headset microphone is enough.
- Confirm consent and create. Give the voice a name and confirm it is your own voice, or that the speaker has given you explicit permission.
- Type a script and generate. Paste your text, generate, listen, and download the MP3.
Getting a natural-sounding result
The quality of a clone depends far more on the sample than on its length. A clean 45-second recording beats a noisy three-minute one. Record somewhere soft-furnished to avoid echo, keep the same distance from the microphone, leave short pauses between sentences, and speak in the tone you want to hear back — a calm narration sample produces a calm narration voice. Avoid music, a TV in the background, or a second person talking.
Where creators use it
- YouTube: fix a mistaken line or voice an entire video without re-recording.
- Multilingual content: publish other-language versions of a video or course in your own voice.
- Courses and training: update a lesson by editing the script instead of booking another session.
- Podcasts and ads: record intros, sponsor reads and pick-ups while you're away from your mic.
Built around consent
A tool that copies voices can be misused, so we built Voice Clone with limits from day one:
- Every voice requires you to confirm it is yours or that the speaker gave explicit permission, and we keep a record of that confirmation.
- Voice Clone is available only on paid Pro and Agency accounts, and the number of voices you can create each month is capped.
- Every file carries an "AI-generated voice" label in its metadata.
- Deleting a voice permanently deletes its sample and voice model.
- Impersonation, fraud and harassment lead to account termination. If you believe your voice was cloned without permission, email audioflowteam@speakscribe.me.
The full rules are in section 4a of our Terms of Service.
Plans and limits
- Pro — $3/month: 3 cloned voices, 50,000 characters of cloned speech per month.
- Agency — $6/month: 10 cloned voices, 100,000 characters of cloned speech per month.
Voice Clone is not part of the Free or Starter plans, but every plan still includes our 400+ stock AI voices.