Captions that are baked into the video image — the big bold words you see on TikTok and Reels — are only one option. For YouTube uploads, online courses, and anything that needs real accessibility, what you actually need is a subtitle file: a separate .srt or .vtt document that rides alongside the video. Viewers can toggle it on and off, pick their language, and search its text. Here’s how to decide which one you need, and how to generate the file itself.

If you already know you need the file, skip ahead and generate an SRT from your video — it’s free, runs in your browser, and hands you a downloadable .srt you can upload anywhere. (If you want the always-visible kind instead, use the auto captions tool for burned-in captions with no watermark.)

Subtitle files vs. burned-in captions: what’s the difference

A subtitle file is a sidecar — a small text file that describes what is said and when, but doesn’t change your video file at all. Burned-in captions are pixels — the words are drawn into the video frames permanently.

That one difference drives everything:

  • Toggleable: sidecar subtitles can be turned on/off by the viewer; burned-in can’t.
  • Multilingual: one video + five .srt files = five languages. With burned-in, you’d re-render the whole video per language.
  • Searchable: YouTube indexes the text of uploaded subtitle files, so your spoken content becomes searchable. Burned-in text is invisible to search.
  • Editable: fix a typo in an .srt by opening it in Notepad. Fix a typo in burned-in captions by re-rendering the video.
  • Platform limits: some platforms (Instagram posts, older players) don’t read sidecar files — burned-in is your only option there.

Rule of thumb: long-form, searchable, accessible content → subtitle file. Short-form social where text is part of the visual design → burned-in.

When you actually need a subtitle FILE

1. YouTube uploads

YouTube lets you upload subtitle files per video and even per language. This matters for three reasons: viewers who are deaf or hard of hearing get accurate captions (YouTube’s auto-captions are getting better but still mangle names, jargon, and accents), non-native speakers can read along, and YouTube’s search can index your actual spoken words. A 20-minute tutorial with an uploaded .srt is discoverable through phrases you said but never wrote in the title or description.

2. Online courses and training

Course platforms — and corporate LMS systems — commonly require real caption files for accessibility compliance. Burned-in captions don’t count toward WCAG requirements because assistive tech can’t parse text baked into pixels.

3. Accessibility (the real kind)

Roughly 1 in 20 people has significant hearing loss. Proper subtitle files let them control font size, contrast, and position through their player’s settings — burned-in captions offer none of that. If your audience includes professionals watching training videos at work, educators, or government clients, subtitle files aren’t a nice-to-have; they’re expected.

4. Multilingual audiences

Say your video is in English but a chunk of your audience speaks Spanish. Translating an .srt file is a text-editing job you can hand to a translator for a fraction of the cost of re-editing video. Upload the translated file as a second subtitle track and YouTube offers it as a language option automatically.

5. Podcasts and talking-head videos repurposed as articles

Many creators publish a video and a blog post version. An .srt file is already 80% of a transcript — strip the timestamps and you have the article draft. (It’s also why subtitle files are handy SEO fuel: the exact words your audience uses.)

When burned-in captions are the right call

Short-form video is the main exception. On TikTok, Reels, and Shorts, captions are a visual element: big, styled, word-by-word, designed to hold attention with sound off. These platforms either don’t accept sidecar files or bury the option so deep nobody uses it. Burned-in is also right when the text is the joke or the hook — kinetic typography that animates with the speech can’t live in an .srt.

Our auto captions tool handles exactly this case: upload a clip, get styled burned-in captions, no watermark, no signup.

How to generate an SRT from your video

The workflow is the same whether you do it manually or with a tool:

  1. Transcribe. Get the spoken words as text. Automatic transcription does in minutes what takes an hour by hand — upload your video to the SRT generator and it transcribes the audio in your browser.
  2. Split into blocks. Each subtitle should be one idea, max two lines, on screen for 1–6 seconds. Viewers read roughly 150–180 words per minute, so keep blocks short — and adjust any auto-split block that lands mid-thought.
  3. Check the timestamps. Scrub through: does each line appear when the words are spoken? Fix any block that’s early or late by more than half a second — that’s the threshold where viewers notice.
  4. Clean the text. Remove filler words (“um”, “uh”) unless they carry meaning, fix proper nouns the transcription missed, and add [music] or [laughter] notes where sound matters for deaf viewers.
  5. Download and upload. Save the .srt, then attach it to your video on YouTube (Subtitles → Add language → Upload file) or hand it to your course platform.

Format gotchas that break subtitle files

An .srt looks simple, and it is — which is why tiny mistakes break it:

  • Timestamp format: SRT uses commas — 00:00:12,500 --> 00:00:15,000. VTT uses periods — 00:00:12.500 --> 00:00:15.000. Mix them up and players reject the file.
  • VTT needs a header: the first line of a .vtt file must literally be WEBVTT, followed by a blank line.
  • Encoding: save as UTF-8. Accented characters and emoji turn into garbage in other encodings.
  • Line length: keep each line under ~42 characters. Long lines get auto-wrapped awkwardly by players and become unreadable on phones.
  • Sequence numbers: SRT blocks must be numbered 1, 2, 3… in order with no gaps. One missing number and some strict players stop reading the file there.

The SRT generator handles all of this automatically — correct numbering, comma timestamps, UTF-8 output — so you can focus on checking the words instead of the syntax.

A real example

Take a 12-minute product tutorial. You upload it to YouTube with an .srt file:

  • A deaf viewer watches with subtitles styled to their preference — bigger text, high contrast.
  • A Spanish-speaking viewer picks the Spanish subtitle track you uploaded last week.
  • Someone Googles the exact error message you said aloud at minute 7 — YouTube indexed your subtitle text, and your video shows up.
  • Next month you turn the video into a help-center article: the .srt becomes the transcript draft in minutes.

None of that happens with burned-in captions. The file is small and free to make — generate yours with the free SRT generator and upload it alongside your next video.