Text to Speech for CapCut: AI Voiceover Workflow Fish Audio
Build CapCut voiceovers that do not sound like the stock voice: write to a word budget, generate in the browser, download the file and drop it on the timeline. Packs from $9.
CapCut made editing approachable, and its built-in text to speech has a sound: millions of videos later, viewers recognize it instantly. The voiceover is the part of a video people hear the most, and it is the cheapest part to upgrade. Generate the voice here, drop the file on the CapCut timeline, and keep the rest of your workflow exactly as it is.
Why generate the voice outside CapCut?
Four reasons, in the order they show up in real projects:
- Voice choice. The library hosts more than 2,000,000 voices, including narration and character registers that stock editor voices do not cover.
- Direction. Emotion tags and pacing descriptions let one line land differently from the next — the text to speech guide covers the syntax.
- Portability. The file you download is not tied to one editor. The same voice can voice a CapCut short on Monday and a long-form edit on Friday without re-casting.
- Rights clarity. Commercial usage rights start at a clear line — the $29 Creator pack — instead of a per-seat subscription.
What does the workflow look like end to end?
- Write to the cut. Budget the script against the video's length (table below), hook first.
- Generate in the browser. No install: open the workspace, paste the script, pick the voice, generate. Listen at normal speed for tone, at double speed for pacing.
- Download the file. Every generation is a file you download and keep — see how downloads work.
- Import into CapCut. Drop the file on an audio track, align the first word to the first cut, and let the rest sit. If captions are in the project, keep them from the same script so they stay in sync.
- Fix by regeneration. When one sentence lands wrong, regenerate that sentence and swap the clip. Nothing else on the timeline moves.
What is the word budget?
Natural speech runs at about 150 words per minute, and one credit covers about 1,000 characters:
| Video length | Words | Characters | Approximate cost |
|---|---|---|---|
| 15 seconds | ~38 | ~225 | a quarter credit |
| 30 seconds | ~75 | ~450 | half a credit |
| 60 seconds | ~150 | ~900 | one credit |
| 5 minutes | ~750 | ~4,500 | 4.5 credits |
| 10 minutes | ~1,500 | ~9,000 | 9 credits |
Write slightly under budget. A read that lands 5% short is invisible after a small trim; a read that lands 5% long forces an awkward speed-up that viewers can hear.
How do you keep a series sounding consistent?
The series voice is the asset, not the individual video:
- Settle on one voice per channel or per character and save it with notes.
- Keep the script as the single source of truth, so fixes are re-generations, not re-records.
- Keep tag conventions simple — one or two delivery tags that recur, for example [excited] in hooks and [whispering] for reveals.
- If the channel posts in more than one language, generate each locale from the same script with the same voice; the TikTok guide covers the localized-posting pattern, and dubbing covers the same idea for finished videos.
What does it cost?
Run the math for a realistic month:
- A daily 30 second short is about 14 credits a month.
- A weekly 10 minute video is about 36 credits a month.
- Both together stay under 6% of the $9 Starter pack.
Monetized channels and client work add the $29 Creator pack for commercial usage rights. Packs are one-time purchases, credits stay valid for 12 months, and there is no subscription to cancel.
Frequently asked questions
Is this a CapCut plugin?
No — it is a two-step workflow. Generate the voiceover in the browser workspace, download the audio file, then import it into CapCut as a normal audio clip. Nothing to install, no plugin to keep updated.
How long should the script be for my video?
Natural speech runs at about 150 words per minute: a 15 second video needs about 38 words, 30 seconds about 75 words, a minute about 150. Write to that budget instead of trimming the read after generation.
How much does a CapCut voiceover cost?
A 30 second read is roughly 450 characters, so about half a credit; a 10 minute video lands near 9 credits. The $9 Starter pack holds 1,000 credits, so months of short-form voiceovers fit inside one pack.
Can I fix one sentence without redoing the whole voiceover?
Yes. Regenerate the single line from the script, download the new file and replace that clip on the timeline. Everything else on the track stays where it was.
Can I monetize videos that use these voices?
Commercial usage rights start with the $29 Creator pack, which covers monetized accounts and brand work. Check the current AI disclosure rules of the platforms you post on before publishing.
Related
- AI voice for TikTok - the short-form posting calendar this workflow feeds.
- YouTube narration - the same pipeline for long-form episodes.
- Text to speech download - what happens to the file after you generate it.