Most Shorts get finished in CapCut. The voiceover should land there as a clean MP3, and captions should come from an SRT you already trust — not from a second transcription pass if you can avoid it.
This guide owns the CapCut packaging loop: generate audio + captions elsewhere, then assemble the edit. It is not the YouTube Studio import click-path, and it is not a “best AI voice” roundup.
The CapCut loop in one sentence
Lock MP3 → build visuals to the voice → drop matching SRT → style for mobile → export. Audio leads; everything else follows.
What you need before CapCut
- A Shorts-length script (speakable lines, not an essay)
- An MP3 voiceover from a locked take
- An SRT file timed to that same take
- B-roll, screen recordings, or stills for the visuals
- Clear file names (tip03-voice.mp3 + tip03-voice.srt)
In Shorts Voice Studio you can generate the MP3 and optional SRT together. Keep each generate under 700 characters; split longer ideas into sections and stitch those MP3s on the CapCut timeline in order.
Generate the pair outside CapCut
- Paste your script in Shorts Voice Studio
- Pick language, voice, and mood
- Turn SRT on before you generate
- Listen on a phone speaker once
- Download MP3 + SRT from the same generate
Do not regenerate mid-edit unless the take is wrong. A new MP3 without a new SRT is how captions drift.
CapCut desktop: audio-first click path
- Open CapCut → New project → set ratio to 9:16
- Import → select tip03-voice.mp3 → drag it onto the primary audio track
- Play once and mark cut points where the tip changes (markers or split on the timeline)
- Import B-roll / screenshots → place clips under the spoken beats (not before the MP3)
- Lower any music track so narration stays readable on a phone speaker
Why audio-first saves time
If you assemble B-roll first, every voice regenerate forces a retime. When the MP3 is the boss, visual cuts snap to words and you stop fighting the timeline.
CapCut: import and style the SRT
Menus differ slightly between CapCut desktop and mobile, but the job is the same: import captions from your .srt file, then style them for a phone screen.
- Desktop: Text / Captions → Import captions (or Auto captions → Import) → choose tip03-voice.srt
- Mobile: Captions → Import → pick the same .srt you downloaded with the MP3
- Confirm the first hook line appears in the opening two seconds
- Fix names, numbers, and brand spellings once in the caption editor
- Set size, weight, and outline so muted viewers can read
- Move captions away from faces, CTAs, and UI you need people to see
- Leave safe margins on the left and right edges
Paste-ready 20-second tip script (under 700 chars)
Use this shape in Studio, enable SRT, then follow the CapCut steps above. Swap the tip line for your niche.
Stop editing Shorts visuals before the voice is locked. Drop your AI voice MP3 on the CapCut timeline first. Cut B-roll to the words, not the other way around. Import the matching SRT from the same generate so captions stay in sync. Export 9:16, watch it muted once, then upload.
Example: 30-second tip Short assembly
- Hook line on screen with first 2 seconds of MP3
- Tip 1 B-roll under the next spoken tip
- Tip 2 screen recording or still under the second tip
- Close CTA card under the last line
- SRT styled once across the whole cut
If you sectioned a longer script into two Studio generates, place MP3 A then MP3 B back-to-back, then import each matching SRT (or merge carefully) so timings stay honest.
Burn-in vs upload later
- Burn captions into the CapCut export for TikTok and Reels-style feeds
- On YouTube, burn in for muted mobile watch, or upload the same SRT in YouTube Studio for an editable track
- One SRT source keeps both paths honest — do not retype captions from scratch
CapCut auto-captions vs your SRT
CapCut auto-captions can be fine for a rough draft. If you already exported SRT with the voiceover, prefer that file. You paid the timing once; re-running auto-captions invites spelling drift on names and product words.
Export checklist before you upload
- 9:16 aspect ratio
- Voice clear on phone speaker (not only headphones)
- Captions readable muted
- No long dead air after the last word
- Music under the tip line, not over it
- Same voice as the rest of the series
Common CapCut mistakes
- Building the whole video before locking the voice take
- Mixing SRT from take A with MP3 from take B
- Letting music drown the offer or tip line
- Leaving auto-captions when you already have a good SRT
- Regenerating voiceovers mid-edit because the script was never spoken aloud once
- Caption boxes covering the product UI you are demoing
Batching several CapCut projects
Generate five MP3 + SRT pairs in one Studio sitting with the same voice. Then open CapCut and assemble one Short after another. Switching tools every video is what makes the week feel long.
FAQ
Do I generate the voiceover inside CapCut?
You can use CapCut’s own TTS, but this loop assumes you generate MP3 + SRT in Shorts Voice Studio (or another TTS), then finish packaging in CapCut. That keeps voice choice and caption timing outside the editor.
Should I burn captions or upload SRT on YouTube?
Burn-in when you also post to TikTok/Reels or want captions guaranteed on muted feeds. Upload SRT in YouTube Studio when you want an editable track without re-exporting. Many creators do both from the same file.
What if my script is longer than 700 characters?
Split into sections in Studio with the same voice and mood. Place the MP3s in order on the CapCut timeline, and keep a matching SRT for each section.
Why do my captions drift from the audio?
Almost always a mismatched pair — old SRT with a new MP3, or a trim on the audio without shifting captions. Re-import the SRT from the same generate as the audio you kept.
Can I reuse this CapCut project for Reels and TikTok?
Yes for 9:16 tip formats. Export once with burned-in captions when those platforms are part of the plan. Adjust cover/thumbnail habits per app.
Lock the voice track first. CapCut gets faster when the audio is the boss.
Turn SRT on before generate, then drop the pair into a 9:16 CapCut project.
Generate MP3 + SRT for CapCut