UtilityNest

Add Subtitles to Video

Embed captions permanently in a video. Load an SRT, SMI or VTT file, preview Korean or English text on the actual frames, fix the timing with an offset, and download an MP4 that keeps its audio.

Works withYouTube logoYouTubeInstagram logoInstagramTikTok logoTikTok

How to use Add Subtitles to Video

  1. 1Drop a video. On a computer it can be any length and up to 50 GB. MP4, MOV, M4V, WebM, MKV and AVI are accepted, and the page checks that the file has a readable video track.
  2. 2Click Choose subtitles and pick an SRT, SMI or VTT file under 2 MB. The page counts how many cues fall inside the clip.
  3. 3If the SMI file holds more than one language, pick the track you want. Enter an offset in seconds: a positive number shows the text later, a negative number earlier.
  4. 4Choose Small, Medium or Large text and place the captions at the bottom or the top. Use Jump to first caption to check the size and position on a real frame.
  5. 5Click Create captioned MP4, wait for the progress to reach 100%, play a captioned section in the result player, then download the MP4.

Check the offset against a known cue

If a caption that should appear at 1.4 seconds appears at 1.0 seconds, enter +0.4 seconds. For broader timing or format changes, use the Subtitle Converter before embedding captions.

When burned-in captions are the right choice

A separate subtitle file only works when the player supports it. Many social feeds, messaging apps, presentation slides and kiosk screens play the video track and nothing else, so a sidecar SRT is simply ignored. Burning the text into the pixels guarantees that every viewer sees it, including people watching with the sound off on a phone. It is also the safest way to hand a captioned clip to someone who will upload it themselves, because there is no second file to lose.

The trade-off is permanence. Viewers cannot switch the captions off, change their size or choose another language. Keep the original video and the subtitle file so you can make a new version later.

What we saw when we tested it

We used a 45-second, 1280 by 720 MP4 at 30 frames per second with a stereo AAC soundtrack (17.4 MB) and an SRT with nine cues that mixed English lines and Korean lines such as 이번 주 신상품을 소개합니다. With an offset of +0.4 seconds, Medium text and bottom placement, the first version of this tool took 84 seconds to render the clip, including loading the encoder. The download was a 16.2 MB MP4 with H.264 video and AAC audio, still 1280 by 720 and 45.0 seconds long. A frame grabbed at the 6-second mark showed the Korean line in white with a dark outline, centred near the bottom, with every syllable intact.

The tool now draws captions with the browser's own video engine. A 40-second 1280 by 720 clip with three captions, one English line long enough to wrap onto two lines and one Korean line, finished in about 12 seconds in headless Chromium, roughly three times faster than real time. Every frame is still decoded, drawn on and encoded again, so a long 4K video takes a while; keep the tab open and the computer awake.

How the captions are drawn

Text is set in a Korean-capable sans-serif font, white with a thin black outline, inset 7% from the left and right edges and 9% from the top or bottom. Medium text is about 5% of the shorter side of the frame, so it scales with the video rather than using a fixed pixel size. Long lines wrap automatically and line breaks from the subtitle file are kept. Styling tags such as italics or colours inside the subtitle file are removed, so every caption looks the same.

The saved video is encoded as H.264 by your browser's built-in video engine, at a slightly higher bitrate than a normal compress so the text edges stay clean. AAC sound is copied as it is; other sound is converted to AAC at 128 kbps. Only the first audio track is kept. Results over about 1.2 GB are written straight to a file on your disk as they are made (Chrome and Edge ask where to save first), so a two-hour recording does not need gigabytes of memory.

Limits to know before you start

  • Length and size: any length, up to 50 GB in Chrome or Edge on a computer, 20 GB in Safari or Firefox and 8 GB on a phone, including 4K. Browsers without a built-in video encoder fall back to a slower method limited to 3 minutes and 100 MB.
  • Cue count: between 1 and 3,000 cues must fall inside the clip after the offset. Cues that end up before the start or after the end are dropped.
  • Formats: SRT, SMI and VTT only. ASS styling is not used, and VTT positioning settings are ignored.
  • Output: always MP4, even if you started with WebM or MKV.

For a longer video, cut the section you need with Trim Video and caption that part.

Common problems and fixes

The page says no valid cues were found. The file is probably not real SRT, SMI or VTT, or its timestamps are malformed. Open it in a text editor and check that each cue has a start and end time, or run it through the Subtitle Converter.

The preview does not play. Some codecs inside MKV or AVI files cannot be decoded by the browser player. You can still export, and the MP4 result plays in the page.

Zero cues in range. The offset pushed every cue outside the clip, or the subtitle file belongs to a longer video. Reset the offset to 0 and check the first timestamp in the file.

Export fails partway. Close other heavy tabs, try a shorter clip, or convert an unusual source to MP4 first with MOV to MP4 or MKV to MP4.

Frequently asked questions

Can viewers hide the captions?

The MP4 contains the captions as visible pixels, so they stay visible wherever the video plays.

Is my video uploaded anywhere?

No. Decoding, drawing the captions and encoding all happen in your browser. The encoder files are downloaded once, but your video and subtitles stay on your device.

Does it handle SMI files with Korean and English tracks?

Yes. When an SMI file defines more than one language class, a Subtitle track menu appears so you can burn in the one you want.

Why is the output a different size from my original?

The video is re-encoded at a quality setting chosen for sharp text, so the file can come out slightly smaller or larger than the source. In our test a 17.4 MB clip became 16.2 MB.

Will odd frame sizes work?

Yes. H.264 needs even dimensions, so a frame that is an odd number of pixels wide or tall is trimmed by one pixel before encoding.

Related guides