How to add captions and subtitles to a video automatically
To add subtitles to a video automatically, upload it to FilmScribe or paste a YouTube link. We transcribe every word with its timing, you fix names and terms, pick one of 235 styles and download the clip with captions burned in for Reels, TikTok and Shorts, or an SRT file for YouTube and your editor.
Step by step
- 01 Upload the video Upload the file to FilmScribe or paste a YouTube link. We transcribe every word with its timing.
- 02 Review the transcript Fix names and technical terms in the text. The fix carries over to the captions on every clip.
- 03 Pick the section Use one of the suggested clips, scored 0 to 100, or mark the start and end in the transcript.
- 04 Pick the style One of 235 styles, with 54 fonts, per-word highlights and automatic keywords.
- 05 Position the captions Drag the captions clear of the areas covered by the profile name, description and app buttons.
- 06 Download and post Download the MP4 with captions burned in for Reels, TikTok and Shorts, or the SRT for YouTube and your editor.
In short
- Automatic captions start with a transcript that has the timing of every word. That’s how FilmScribe builds them, in 90+ languages.
- There are two kinds of captions: open (burned into the picture) and closed (an SRT file viewers turn on and off). FilmScribe delivers both.
- Word-by-word animated captions are the norm on Reels, TikTok and Shorts, because most people start watching with the sound off.
- Proofreading is a must with any tool. In FilmScribe, you fix the transcript once and the fix applies to every clip.
- Position matters as much as style. Captions under the app’s buttons are captions nobody reads.
Burned-in or SRT: decide first
The first choice is the kind of caption, because each destination calls for one.
| Aspect | Open captions (burned in) | Closed captions (SRT file) |
|---|---|---|
| How it works | Part of the video image | Separate file, synced by the player |
| Can viewers turn it off? | No | Yes |
| Style | Anything: animation, color, boxes | Set by the player |
| Where to use it | Reels, TikTok, Shorts, stories, ads | YouTube, courses, long videos, editors |
| Search and accessibility | Not read as text | The platform reads and indexes the text |
In FilmScribe, both come from the same reviewed transcript: the MP4 clip with captions burned in and the SRT file. On YouTube, do both: the long video gets an SRT, the vertical clips get burned-in captions.
How to add captions automatically in FilmScribe
1. Upload the video or paste a link
Upload the file (MP4, MOV, MKV, WEBM, AVI and more) or paste a YouTube link, including a livestream, which is processed once the replay is available. If only part of it matters, pick the range before processing: only that range is processed and counts toward your plan’s hours. We transcribe in 90+ languages, detect the language automatically and separate the speakers.
2. Review the transcript
We transcribe word by word, with the timing of every word. Use the transcript search to find each name and fix it right in the text. The fix carries over to the clip captions, so you review once, on the whole video, not clip by clip.
3. Pick the section
For vertical video, FilmScribe suggests up to 60 clips per video, each with a score from 0 to 100 and the type of moment. You can also mark the section in the transcript and fine-tune the start and end word by word. The clip comes out in 9:16 with the camera following whoever is speaking, in 1:1, 4:5, 16:9 or the original format.
4. Pick the style
There are 235 styles, from classic TV to word-by-word animation: viral, karaoke, boxes, neon, 3D, retro, comic, news, gamer, podcast and more, with 54 fonts. On Studio and Agency, you can also upload your own font as TTF or OTF. For a complete look, start from one of the 48 templates, with the clip title built into the design.
5. Fine-tune the details
- Per-word style: colors, boxes and size on the words you choose.
- Automatic keyword highlights, by color, box or underline.
- Free positioning: drag and resize the captions right on the video.
- Line breaks, a capital at the start of each word and punctuation removal.
- Hide sounds tagged in the transcript, like [music], [laughter] and [applause].
The FilmScribe preview shows exactly what ends up in the video.
6. Download what the destination needs
- MP4 with captions burned in, ready for Reels, TikTok and Shorts, with no watermark on paid plans and the original recording quality.
- SRT with TV, Netflix and YouTube presets and your project’s frame rate.
- FCPXML with the captions for Final Cut.
Burned-in captions are for video. If you uploaded audio only, you still download the transcript as SRT, TXT or PDF. A captioned clip can run up to 15 minutes on Pro and up to 25 minutes on Studio and Agency. See every detail in animated captions.
How word-by-word animated captions work
Classic captions work in sentence blocks: they appear, stay a few seconds and disappear. Animated captions need the timing of every word to know when to highlight the next one.
A traditional SRT block:
12
00:01:04,200 --> 00:01:07,480
God doesn't give up on people
who already gave up on themselves.The same line with the timing of every word, which is how the FilmScribe transcript arrives:
00:01:04,200 God
00:01:04,480 doesn't
00:01:04,790 give
00:01:04,980 up
00:01:05,150 on
00:01:05,290 people
00:01:05,840 who
00:01:06,010 already
00:01:06,390 gave
00:01:06,620 up
00:01:06,790 on
00:01:06,900 themselves.With that timing, the style lights up “give up” exactly when the pastor says it: it changes color, scales up or gets a box behind it. Without it, every animation drifts off the voice.
Where the captions go: 3 destinations
- Reels, TikTok and Shorts. Download the clip from FilmScribe with captions burned in and post it. Each platform’s quirks are in the guide on how to add captions to Reels and TikTok.
- A long video on YouTube. Download the SRT with the YouTube preset and upload it in the video’s subtitles section. Viewers turn it on and off, and the text is readable for search.
- Your editor. Take the SRT or FCPXML and finish the styling there. On Studio, Agency and their trial, the cut timeline also exports as .xml for Premiere Pro and DaVinci Resolve or .fcpxml for Final Cut Pro. The guide on how to create an SRT file covers the file in detail.
In all three cases, the order is the same: timed transcript, proofreading, style and position. Only the output format changes.
Captions in another language
If your audience speaks another language, Studio and Agency translate captions into 9 languages: English, Spanish, French, German, Italian, Portuguese, Japanese, Korean and Simplified Chinese. Translated captions can be burned into the clip itself in all of them except Japanese, Korean and Chinese, which come out as files only. See translation and the guide on how to translate subtitles.
Readability rules
Broadcasters and streaming services follow caption guidelines with similar numbers. Use these as a starting point for SRT:
| Rule | Common reference | Why |
|---|---|---|
| Reading speed | up to 20 characters per second | Faster than that and viewers fall behind |
| Characters per line | up to 42 for TV and streaming | Long lines tire the eye and cover the picture |
| Lines per caption | 2 at most | Three lines cover the video |
| Line breaks | At natural phrase boundaries | “on people / who already” reads better than “on / people who” |
Vertical video with animated captions follows a different pattern: short blocks of 2 to 5 words on one line, because the screen is narrow and the type is big. To check an SRT, use the subtitle reading speed checker.
Font, style and position
- A bold sans serif. Wide, heavy letters read well on a small screen. Avoid thin and script fonts.
- Outline, shadow or box. Captions cross light and dark backgrounds in the same video.
- Highlight with color, not another font. In FilmScribe, per-word style changes color, box or size without swapping the font.
- A style that fits the content. Sermons call for something clean; podcasts can take more color; weddings need a light touch; lessons need clarity.
- The same identity on every video. Same style and colors across the channel.
On Reels and TikTok, the bottom of the screen is covered by the profile name, description and music, and the right edge holds the buttons. Drag the captions clear of those areas and check with the Reels and TikTok safe zone tool.
Auto captions in CapCut, Premiere and DaVinci
These editors also generate captions from audio. CapCut makes auto captions in a few taps; Premiere Pro does it from the sequence transcript; DaVinci Resolve keeps captions from audio in the paid version, and in the free one you import an SRT. For a short one-off video, that’s fine.
The work shows up in the routine: in every project you proofread again, reapply the style and, for word-by-word effects, animate line by line. With FilmScribe, proofreading happens once in the transcript and the clips come out captioned in the same style. If you finish in an editor, take the SRT, and the guide on animated captions in Premiere and DaVinci shows how to drop the finished clip onto the timeline.
Proofreading checklist before you post
Do this once in the FilmScribe transcript, on the whole video:
- Names. Pastor, guest, church, company, couple. Search for each one in the transcript and check the spelling.
- Numbers and dates. Pick a format and stick with it. Digits read faster on a small screen.
- Sound-alikes. “Their” and “there”, “then” and “than”, “affect” and “effect”.
- Punctuation. On animated captions you can drop it for a cleaner look. On SRT for YouTube, keep it.
- First and last caption. The first one lands with the hook; the last one can’t vanish before the final word.
- Sensitive words. Decide up front whether they appear spelled out or not.
Edge cases
- Music and worship in the middle of the video. Hide sounds in the captions and trim the clip’s start and end to keep only speech.
- Two people talking at once. Overlapping speech makes captions messy in any tool. Prefer a section where one person finishes before the other starts.
- A very long video. For a three-hour livestream, download the SRT of the whole video for YouTube and burn captions only into the clips. Picking the range before processing keeps you from spending hours on the parts you don’t need.
- Captions out of sync in the editor. It’s almost always a frame rate mismatch. Download the SRT at the project’s rate or fix it with the subtitle sync tool.