Most social video is watched with the sound off — on mute in a feed, in a waiting room, next to a sleeping baby. Captions aren’t an accessibility bonus anymore; they’re the primary way a huge share of your audience consumes the content. And on TikTok, Reels and Shorts, the reliable way to guarantee captions appear is to burn them in: render the text permanently into the video frames.

Our free subtitle burner takes your video plus an SRT file and produces a captioned MP4 in your browser. But burning is permanent — which means the subtitle file and the styling have to be right before you commit. Here’s the full playbook.

Burned-in vs. soft subtitles: the real tradeoff

Burned-in (hardcoded):

  • ✅ Displays everywhere, no player support needed, no toggle — the viewer can’t miss them
  • ✅ You control the exact styling: font, size, color, position
  • ✅ Works on every social platform identically
  • ❌ Permanent — typos are baked in; fix the SRT and re-burn
  • ❌ Can’t be turned off by viewers who don’t want them
  • ❌ Not indexed as text by platforms (YouTube can’t “read” burned captions for SEO)
  • ❌ One language per export — multilingual means multiple renders

Soft subtitles (separate SRT/VTT file or platform captions):

  • ✅ Toggleable, multi-language, platform-indexed
  • ❌ Depend on player support — inconsistent across social apps
  • ❌ Styling controlled by the platform, not you

The rule: burn for social feeds (sound-off scrolling, no reliable caption support), go soft for YouTube and long-form (SEO value, viewer choice, multiple languages). Many creators do both: burned captions on the TikTok/Reels cut, uploaded SRT on the YouTube version. Our caption guide covers the soft-subtitle side.

Get the SRT right first (burning magnifies errors)

Because burned captions are permanent, the subtitle file deserves a proofreading pass:

  • Timing. Each caption should appear exactly when the words are spoken — not a half-second early or late. Viewers forgive a lot, but desynced captions feel broken. Our subtitle sync guide covers fixing drift.
  • Line length. Max ~32–42 characters per line, max 2 lines per caption. Longer lines force tiny fonts or overflow the frame.
  • Reading speed. A caption should stay up long enough to read comfortably — roughly 150–180 words per minute. If your speaker is fast, split into more captions rather than flashing walls of text.
  • No auto-caption garbage. AI transcription is a starting point, not a finished product. “Their” vs “there,” missing punctuation, and misheard names all get burned in permanently if you don’t proofread. Read the whole SRT once before burning.

Real example. A creator burns auto-generated captions into a 60-second product video without proofreading. The product name “KeraGlow” is transcribed as “carrot glow” in 4 captions. The video gets 200K views — with the wrong product name permanently embedded in every frame. The fix (correct SRT, re-burn, re-upload) costs the video its momentum. Ten minutes of proofreading would have prevented it.

Styling rules for burned captions

Position: lower third, but above platform UI. Centered horizontally, sitting in the lower third of the frame. Critical: on TikTok/Reels/Shorts, the bottom ~15% is covered by captions, progress bars and buttons — place your burned subtitles above that zone or they’ll be hidden behind the platform’s own UI. Test with a real upload before committing to a template.

Font: bold, sans-serif, high contrast. The classics are classics for a reason — heavy weights of Inter, Roboto, Arial Black or similar. Thin elegant fonts are unreadable at caption sizes on phone screens. White text with a black outline or semi-transparent black background box is the maximum-readability standard.

Size: big enough to read at arm’s length. On a phone held normally, caption text should be roughly the size of the platform’s own UI text or larger. When in doubt, go bigger — undersized captions are the most common burned-subtitle failure.

Word-by-word highlighting (karaoke style). The TikTok-native style where each spoken word highlights in sync. It’s high-effort but measurably engaging — the moving highlight pulls the eye along and increases completion. If your burner supports it, this is the premium option for talking-head content.

Emphasis styling. Bold or color key words (prices, product names, CTAs). Sparingly — if everything is emphasized, nothing is.

What NOT to burn

  • The entire transcript as a paragraph. Captions are timed to speech; a wall of static text is a thumbnail, not a subtitle.
  • Captions over important visuals. If the lower third contains the product demo, move captions up or shrink the demo. Plan caption space when framing the shot.
  • Speaker labels on single-speaker content. “[SPEAKER]” tags from transcription tools look broken when there’s one person talking. Strip them.
  • Sound-effect descriptions on social cuts. “[upbeat music playing]” is correct for accessibility on YouTube; on a 30-second Reel it’s clutter. Know your platform.

The workflow

  1. Generate or write your SRT (our SRT generator or auto-captions to start).
  2. Proofread the full file — timing, spelling, line lengths. This is the step that matters most.
  3. Decide styling: position above platform UI, bold sans-serif, high contrast.
  4. Open the subtitle burner — local processing, nothing uploads.
  5. Burn, then watch the entire export once at phone size. Check every caption for overlap with UI zones and typos.
  6. Keep the SRT file — you’ll need it for the YouTube version’s soft captions and for any re-burn.

Burned-in subtitles are a commitment — that’s their strength and their risk. The platforms where they matter most (sound-off social feeds) reward them enormously: captioned videos consistently outperform uncaptioned ones in retention precisely because the sound-off majority can actually follow along. Do the proofreading, respect the platform UI zones, and burn with confidence.