AI transcription + voice-over studio
Catalog voices only · No cloningUpload audio or video, review an editable transcript, then translate or re-voice it with 56 catalog voices — 30of them premium, on paid plans. Or start with a script and create the read from scratch.
Live TTS demo needs no account · Transcription starts after sign-in
Start with what you have
See the recording workflow or try the live text studio.
off-script-interview.mp3
2 speakers · English · 00:31
Sample source recording · Play it, then compare the raw and cleaned transcript.
Transcript
Click a paragraph to jump in the recording
Ready to re-voice
Maya · Everyday Host / Dev · Commentary
The reviewed transcript opens in the editor with speaker turns intact, ready for voice assignment and generation.
Interactive example · Uploads stay private to your account
Try your recordingSeparate voices into an editable script when diarization is available.
Check exactly what was heard, then review the polished version.
Cast catalog voices, create new audio, and download timed captions.
One connected workflow
Transcription is not a dead-end text export. The transcript becomes the script for the same editor, voices, and export tools.
Upload audio or video. ToneCraft extracts the speech and keeps the recording beside the transcript while you review.
Compare raw and cleaned text, check speaker turns, and edit names or lines before anything is re-voiced.
Assign catalog voices, translate for spoken delivery if needed, then export the voice-over and SRT captions.
Speech-to-text use cases
Recover scripts, edit recordings, create captions, or prepare a translated re-voice. The transcript is a starting point you can keep working with.
Podcast and interview edits
Separate speaker turns, compare raw and cleaned text, then assign a catalog voice to each role for the new read.
You leave with
Multi-speaker script + voice-over
Video captions
Extract the speech, review it beside the recording, and export SRT captions without rebuilding the script by hand.
You leave with
Editable transcript + SRT
Spoken localization
Clean the transcript, translate the script, and produce a new voice-over with catalog voices in a supported language.
You leave with
Translated script + voice-over
Meetings and calls
Upload a recording or import a Zoom or Teams transcript, then clean names, filler, and speaker labels before reuse.
You leave with
Clean, editable transcript
Voice memos and rough takes
Capture the thought in your own voice, revise the transcribed wording, then cast a catalog voice for the final performance.
You leave with
Polished script + narration
Recover a lost script
Recover the words behind an old episode, lesson, or voice-over so you can update lines without starting from zero.
You leave with
Recovered script ready to edit
Listen first
Studio-quality voices for every kind of read — tap one to hear it.
Every voice speaks all 20 languages. The 30 premium voices are on paid plans.
Hear all 56 voices →A deliberate boundary
ToneCraftis built for producing narration, not copying someone's identity. The catalog is fixed, previewable, and governed by clear use rules.
Catalog voices by design
Choose a purpose-built voice for hosts, lessons, stories, ads, and commentary. No uploaded identity is needed.
A complete publishing workflow
Direct multiple speakers, generate a level-matched track, and export MP3, WAV, or SRT from the same project.
Rules you can inspect
No scams, robocalls, deepfake impersonation, or celebrity imitation. Acceptable Use · Refunds · Contact
What an hour costs
ToneCraft
≈ $2.33
$7/mo · ≈ 3 hours
ElevenLabs (Creator)
≈ $18
$22/mo · ≈ 1.2 hours
Murf (Creator)
$14.50
$29/mo · 2 hours
Entry paid plan on each side. Hours are derived at 100,000 characters per spoken hour, except where a provider publishes time directly. Third-party prices checked July 13, 2026.
ToneCraft is AI text-to-speech for creators: turn a script into studio-quality voice-over for podcasts, audiobooks, courses, ads, and video narration. You pick from our catalog of voices — we do not offer voice cloning of real people or deepfake impersonation tools.
Every plan includes a monthly character allowance — roughly 1,000 characters per minute of speech. A 10-minute narration is about 9,000–10,000 characters. Regenerating an unchanged paragraph is free: we cache identical audio and never charge twice for the same text.
You own what you generate — our speech provider assigns output rights to the generating account, and we pass them through to you in full, for your own podcasts, audiobooks, videos and courses. Free-tier audio includes a short spoken "ToneCraft" outro. One rule carries through from those terms: don't present generated narration as a human recording where that would mislead your listeners.
20 languages through 56 studio-quality multilingual voices — 26 on every plan and 30 premium voices on paid plans — with per-word pronunciation control via your dictionary.
Your next project
Recover the script, review the speakers, then turn it into a voice-over you can edit and publish.
Catalog voices only · No voice cloning · Audio and captions stay separate from source video
