Manage by segment
Paste a script and it is split automatically by your per-generation limit. Split, merge or add segments any time.
Voice projects
Move from script to finished audio. Regenerate only the sections that change.
Loading projects…
- Long-form Scripts
- Auto Segmenting
- Batch Generation
- Partial Redo
- Continuous Playback
- Multi-format Export
- Long-form Scripts
- Auto Segmenting
- Batch Generation
- Partial Redo
- Continuous Playback
- Multi-format Export
What is an AI voiceover project?
Long-form text to speech as a work you can keep editing, not a one-off file
Regular text to speech handles a few hundred characters at a time, and each result is a standalone file. Change one sentence and you usually redo the whole thing. Voice Projects fix that for long-form voiceover: your script is split into sections that are generated, edited and versioned on their own, then merged in order into one finished audio file.
Projects support multiple speakers. Paste a script written as "Host: line" or "[Narrator] text" and speakers are detected automatically; you just pick a voice for each. If your script isn't labeled, paid plans can detect speakers with AI. Write [pause 800ms] to control pauses, or @Speaker {speed=1.2} to change speed and mode for one section.
Paste your script or import TXT, Markdown, Word, CSV, Excel or SRT/VTT subtitles. You'll see total characters and credits before generating, and editing a section only regenerates that section; everything else and your selected takes stay put.
When you're done, play the whole project back to back and export one merged MP3, a ZIP of section MP3s (with custom file names and a manifest), or the script. Ideal for audiobooks, kids' stories, course narration, podcasts, game lines and video voiceovers.
- Max sections per project (Premium)
- 3,000 sections
- Max text per project (Premium)
- 300k chars
- Speakers per project (by plan)
- Up to 20
- Export formats
- MP3 / ZIP / SRT / VTT
Long-form AI voiceover, section by section
Audiobooks, courses, explainer videos and multi-speaker scripts, all made in parts
Redo only what changed
Edit a segment and regenerate just that one. Everything else stays as is, saving time and credits.
Export in one go
Play the whole project back to back, then export a merged MP3, a ZIP of segments, or the script.
What Voice Projects are for
Any long content made in parts and revised often
Audiobooks and kids' stories
Paste chapter by chapter, voice narrator and characters separately, and fix a typo by regenerating one line.
Courses and training
Turn notes into a full lesson voiceover and redo only changed sections when the material updates.
Podcasts and conversations
Give the host and guest their own voices, control pauses between turns and export the full episode.
Video voiceovers
Split narration by shot, adjust speed per section and export separate files to align in your editor.
Game lines and prompts
Import lines and file names from a spreadsheet, generate in bulk and export a ZIP named the way your project expects.
Multilingual editions
Keep the same structure, swap in a translated script and produce another edition with the same voices.
An AI voiceover project in three steps
From script to finished, deliverable audio
- 1
Create a project
Paste your script and choose the default voice, language and speed.
- 2
Generate in batch
Generate every segment at once, then fine-tune the ones you don't like.
- 3
Review and export
Play it all back and export as MP3, ZIP or script.
6 tips for long-form voiceover projects
Save credits and avoid rework
- 01
Preview a section or two first
Check the voice, speed and pauses before batch generating, so you don't discover you need a different voice after the whole script is done.
- 02
One idea per section
Shorter sections are easier to revise. A blank line starts a new section, and long sentences can be split.
- 03
Use pauses for structure
Leave [pause 1.5s] between chapters; 300–600 ms between dialogue turns sounds most natural.
- 04
Keep speaker names consistent
Use the same name for a speaker throughout, like always "Host", not a mix of "Host" and "Presenter".
- 05
Use the script view for big edits
For heavy rewrites, switch to the script view. Before saving you'll see what changed, and untouched sections keep their audio.
- 06
Export for how you deliver
Publish with the full MP3; choose the section ZIP to align in an editor or to deliver lines as separate files.
Voice Projects vs single text to speech
When to start a project
| AI voiceover project | Single text to speech | |
|---|---|---|
| Text length | Up to 300k characters per project, by plan | 500 characters per generation (1,000 on Premium) |
| Multiple speakers | Detected automatically, one voice each | One voice per generation |
| Revisions | Regenerate only changed sections, keep old takes | Regenerate the whole text |
| Pauses and settings | Per-section pause, speed and mode | One setting for everything |
| Export | Full MP3, section ZIP, script | One audio file |
Voice Project specs
Project limits by plan
- Sections per project
- 100 / 500 / 1,500 / 3,000 by plan
- Characters per project
- 20k / 100k / 200k / 300k by plan
- Characters per section
- 500 (1,000 on Premium); longer text is split automatically
- Speakers
- Free 2, Starter 4, Standard 8, Premium 20 (including narrator)
- Import formats
- TXT, Markdown, Word (.docx), CSV/TSV, Excel (.xlsx), SRT/VTT
- Pauses
- 0–5 seconds, per section or in bulk
- AI speaker detection
- Starter and above, 1 credit per 10 characters
- Export formats
- Full MP3, section MP3 ZIP (with manifest), SRT / VTT subtitles, TXT script
- Generation pricing
- 1 credit per Chinese character or 3 English characters, only for sections you generate
Voice Projects FAQ
Sections, speakers, credits and export
Your script is split automatically by your plan's per-generation character limit. You can then split, merge or add segments freely.
Related tools
Voices and audio for your projects
Start your first project
Paste a long script and get complete voice-over audio in minutes.