# Add burned-in subtitles to a video for TikTok

> Markdown version of https://stella.sh/how/video/burn-subtitles-into-video-for-tiktok for agents and LLMs.
> Burn subtitles into a video so they show on TikTok, Reels and Shorts: transcribe, fix the text, and render with free tools, or let Stella (stella.sh) do it.

Burned-in subtitles are drawn into the video pictures themselves, so they show everywhere, even with captions turned off. Make a timed subtitle file from the speech, correct it, then render it onto the video with a free tool like CapCut or ffmpeg, keeping the text above TikTok's buttons. Stella (stella.sh) does it on your computer: it transcribes the clip, fits a widescreen video to TikTok's 9:16 frame, and saves a captioned copy next to the original.

- Checked: 10 November 2026
- Stella does it on: Windows, Mac, Linux

## Stella doing it

Recorded 10 November 2026 on Linux (Omarchy) with Stella’s own models (GPT-6 Luna).
Result: A 1080×1920 copy of a 39-second widescreen interview, cropped to portrait with 15 burned-in caption lines. Transcribed again independently afterwards, every caption starts within 0.14 s of the speech. The original file was unchanged. Trimmed: a two-minute wait before the message sent (the new test account was still signing in) is cut, and the waits are sped up.

Video (1 min 24 s): https://stella.sh/media/proofs/video/burn-subtitles-into-video-for-tiktok/proof.mp4

Transcript:

- 0:00 The prompt, typed into Stella on a Linux computer.
- 0:05 Sent. Stella replies that it will transcribe the interview, add TikTok-friendly subtitles and export a copy without changing the original.
- 0:10 Transcribing and timing the captions, sped up 20×.
- 0:17 Stella asks how to fit the 16:9 interview into TikTok's 9:16 frame: crop to portrait, or keep the full frame on a blurred background.
- 0:27 “Crop to portrait” is picked.
- 0:31 Cropping, scaling and burning in the captions, sped up 20×.
- 0:39 Done: timed English subtitles and a TikTok-ready portrait version with bold, outlined text. The original is unchanged.
- 0:56 The finished video playing in Stella, captions burned in.

What Stella does:

1. Transcribes the speech with word-level timings (in the recording, with Stella's own transcription, which comes with paid plans)
2. Checks the sound and picture line up; in the recording it found the audio started half a second late and corrected for it
3. Asks how to fit a widescreen clip into TikTok's 9:16 frame: crop to portrait, or keep the full frame on a blurred background
4. Writes a subtitle file and burns it in with ffmpeg: white bold text with a dark outline, two lines at most, above TikTok's caption area
5. Saves a new 1080×1920 video next to the original, which it leaves untouched

## The prompt to paste

```text
Add subtitles to interview.mp4 and burn them in so I can post it on TikTok.
```

## Do it yourself

1. **Easiest: CapCut's auto captions.** In CapCut (desktop or phone), import the clip, choose Text, then Auto captions, and pick the spoken language. Read through every line and fix mistakes, choose a style, and export. The captions are burned in on export.
2. **Free and offline: make the .srt with Whisper.** If you'd rather not upload the clip, install OpenAI's free [Whisper](https://github.com/openai/whisper) (or the faster whisper.cpp) and run `whisper clip.mp4 --model small --output_format srt`. You get `clip.srt`, a plain text file with numbered, timed lines.
3. **Correct the subtitle text.** Open the .srt in any text editor and fix names, slang and misheard words. Keep the numbers and the `00:00:01,200 --> 00:00:03,900` timing lines exactly as they are. Shorter lines read better on a phone: aim for under 30 characters.
4. **Burn them in with ffmpeg.** Run `ffmpeg -i clip.mp4 -vf "subtitles=clip.srt:force_style='Fontsize=16,Bold=1,Outline=2,Alignment=2,MarginV=90'" -c:a copy clip-captioned.mp4`. `MarginV` lifts the text off the bottom edge; raise it if TikTok's caption and buttons cover your words.
5. **Check it on a phone before posting.** Send the new file to your phone and play it full screen. TikTok puts its buttons on the right and the description along the bottom, so the safe area is roughly the middle two-thirds of a 1080×1920 frame.

## Where Stella stops

- Stella doesn't post to TikTok or any other account for you. You upload the finished file yourself.
- Its own transcription comes with paid Stella plans. On the free path (your ChatGPT or Claude plan) it can't use it, so hand it a subtitle file or have it set up Whisper on your computer, which this recording doesn't show.
- Transcription is good but not perfect, especially over loud music, heavy accents or several people talking at once. Add “show me the text before you burn it in” to the prompt to check the words first.
- Rendering takes as long as your computer takes. A one-minute clip is quick; an hour-long video on an older laptop can take a while.
- Animated word-by-word caption templates are CapCut's strength. Stella makes clean, readable captions, not CapCut's full template library.

## Questions

**What's the difference between burned-in and closed captions?**
Burned-in (open, or hardcoded) captions are part of the picture and always show. Closed captions are a separate track the viewer can turn on or off. TikTok and Reels viewers mostly watch muted, so creators burn them in.

**Does TikTok add captions automatically?**
TikTok can generate auto captions that viewers can switch on, but they're small, styled by TikTok, and often off by default. Burning your own in means everyone sees them, in your style.

**Where should captions go on a TikTok video?**
In the lower-middle of the frame, above the description text and left of the like, comment and share buttons. Keep a clear margin of about 15% at the bottom and on the right.

**Should I cut the pauses before or after adding subtitles?**
Before. Cutting moves every line's timing, so captions made first drift out of sync. [Cut the silences out of the video](https://stella.sh/how/video/cut-silences-from-video) first, then caption the shorter file.

**Can I add subtitles in another language?**
Yes. Transcribe in the spoken language, translate the .srt (Stella can translate it and keep the timings), then burn in the translated file the same way.

**Is Stella free for this?**
Partly. Stella itself is free with the ChatGPT or Claude plan you already have, and so are the ffmpeg steps. The transcription in the recording used Stella's own transcription model, which comes with paid Stella plans (from Plus, $20 a month); new accounts get a small free trial of Stella's own models. On the free path, give Stella a subtitle file, or ask it to set up the free, open-source Whisper on your computer and transcribe with that.

**Can I control what Stella changes on my computer?**
Yes. Stella works on Windows, Mac and Linux with your own files and shows each step in the chat, so you can watch it and stop it at any point. Tell it what it may not do, as the prompts on these pages do (“don't delete anything without asking”). In the Downloads recording it deleted nothing until it was told to go ahead. It also refuses commands that would wipe a whole disk, your home folder or system folders.

Sources: [Whisper on GitHub](https://github.com/openai/whisper), [ffmpeg subtitles filter](https://ffmpeg.org/ffmpeg-filters.html#subtitles-1), [CapCut auto captions](https://www.capcut.com/tools/auto-caption-generator).

## Related problems

- [Cut silences out of a video automatically](https://stella.sh/how/video/cut-silences-from-video.md)
- [Compress a video for Discord](https://stella.sh/how/video/compress-video-for-discord.md)
- [Organize a messy Downloads folder](https://stella.sh/how/pc-cleanup/organize-downloads-folder.md)
- [Transcribe voice memos to text](https://stella.sh/how/audio/transcribe-voice-memos-to-text.md)
- [Create an SRT subtitle file from a video](https://stella.sh/how/video/make-srt-subtitle-file.md)
- [Convert a horizontal video to vertical, keeping you in shot](https://stella.sh/how/video/crop-video-to-vertical.md)

Get Stella: https://stella.sh/download. Free with the ChatGPT or Claude plan you already have.
