How to Add Subtitles to YouTube Videos: SRT, Burned-In Captions and Translations

Last updated: October 2026

A lot of vlogs are watched with the sound off, on a train, at work, or late at night. If your words only exist as audio, those viewers get moving pictures and no story. Captions fix that, but only if you pick the right kind for each platform and clean them up properly. This guide covers SRT files versus burned-in captions, fixing YouTube auto-captions, translated tracks, styling, vertical safe zones, transcripts and audio description.

Why so many viewers read your vlog instead of hearing it

Start with access. Deaf and hard-of-hearing viewers rely on captions for dialogue, and good captions also name key sounds, like music starting or a door slamming. That is the difference between following a scene and guessing at it. Captions belong in the routine from the first upload, as our guide on how to start a YouTube channel covers.

Then there is silent viewing. Phones default to muted playback in feeds, and people watch in offices, buses and beds. A viewer who lands on your vlog with no sound gives you about two seconds. Clear captions can hold them. A silent face talking cannot. Captions also help viewers who miss fast speech, accents and slang. They matter even more for faceless, text-led formats, where on-screen words carry the whole story, as in our guide to 7 vlog setups without showing your face. Hold the rule: if a line matters when spoken, it matters in text.

SRT file or burned-in text: picking per platform

An SRT file is a small text file that sits next to the video. Each block holds a number, a start and end time, and one or two lines of text. The platform reads it and draws the words over the picture. Because the text stays separate, viewers can switch it off, change its size and color, and pick another language. YouTube can also read it for search and translation.

Burned-in captions, sometimes called open captions, are part of the pixels. You add them in your editor, export once, and the words appear on every play, exactly as styled. Nobody can turn them off or enlarge them. That permanence suits short vertical clips. On a 20 minute video it is a weakness, because a viewer who finds the text distracting has no way out.

You can do both: a cleaned SRT for the main video, and a captioned cut for short clips pulled from it. Our YouTube SEO basics for vloggers guide explains how readable text supports discovery alongside titles and descriptions.

Four ways to add subtitles in YouTube Studio

Every route starts in the same place. Open YouTube Studio, choose Subtitles in the left menu, and click the video. If the video language is not set yet, Studio asks you to pick it first, because every caption track hangs off that language. You can also add captions during the upload itself, in the video elements step, so a video never has to go live without them. From the Subtitles screen, pick your route:

• Upload a file. Choose Upload file, then pick With timing or Without timing. With timing suits a finished caption file that already carries time codes, such as an SRT exported from your editor. Without timing suits a plain transcript: you paste or upload the words and YouTube matches them to the speech, setting the timings for you.
• Transcribe and auto-sync. Type or paste the full text of what is said, then let YouTube sync it to the audio. This works well if you filmed from a script or notes, because the words already exist and only the timing is missing.
• Type manually. Build the track block by block in the caption editor, playing the video as you go. The editor can pause while you type, and you set each block’s place on the timeline yourself. Slow on a long video, but you control every word and cue, which makes it a fair choice for a short or mostly silent clip.
• Duplicate and edit the auto track. Take the timing YouTube already worked out and fix the text on a copy. The next section walks through this pass in full, because it is the route most vloggers should take.

On file formats, YouTube accepts several caption file types, including SRT, SBV, SUB, LRC and CAP. SRT is the safe default. Every editor exports it, every platform reads it, and it opens in a plain text editor when you need to fix one line by hand.

Pick the route by trading time against accuracy. Typing everything yourself is the slowest and the most exact. Fixing the auto track is the middle ground most channels settle into. If you have neither the time nor the patience, captioning services take the video and return a timed file, but you still need to spot check names and terms before you publish, because no outside transcriber knows your channel’s vocabulary. Whichever route you choose, the goal is the same: a track you have read with your own eyes before viewers rely on it.

Fixing YouTube auto-captions, step by step

YouTube creates auto-captions from speech recognition. The words are often close. The punctuation and speaker breaks are often wrong, and a wrong name can change a line. Treat the track as raw material:

One catch first: some videos never get an auto track. Speech recognition struggles when a long music intro plays before anyone talks, when several people speak over each other, when the audio is muffled, or when the spoken language is one YouTube cannot caption automatically. If no auto track appears next to your video, skip this section’s first step and use one of the other routes above: upload a file or auto-sync a transcript instead.

1. Duplicate the track. In Studio, open the video, choose Subtitles, and duplicate and edit the auto track. You keep its timing and fix the text on the copy.
2. Fix punctuation and sentence breaks. Auto-captions run sentences together and drop question marks. Break lines where a thought ends, and add periods, commas and question marks so each caption reads like written English.
3. Fix names, places and your regular terms. Search for your channel name, camera names, towns and people. Speech recognition mangles proper nouns first, and those are the errors viewers notice.
4. Check numbers and units. Distances, times and settings like aperture values are easy to mishear as other digits. Confirm each one against what you said on camera.
5. Repair the timing. Drag blocks so text appears as the words start, not a second later. Keep each block up roughly one to seven seconds, and split blocks that cover two shots or speakers.
6. Add sound cues where they carry meaning. Short notes like [music] or [laughter] help deaf viewers and cost seconds. Skip noise that changes nothing.
7. Read it cold, with the sound off. Play the video muted and read only the captions. If you can follow the whole story, the track works. If you stall anywhere, fix that block.

Two editor habits make the pass quicker. Turn on the setting that pauses the video while you type, so the audio never runs ahead of your hands. And if the block view feels fiddly, switch to editing the track as plain text, fix the words there, then return to the timed view for step 5. You can save a draft and come back later at any point. Nothing goes live until you publish the track.

Once the track is clean, download it as an SRT and keep it with your project files. It becomes your source for translations and your transcript. The pass takes about 15 minutes on a typical vlog, and it fits at the end of the workflow in our guide on how to edit vlogs faster.

Translated tracks and localized titles that open new audiences

Start from your cleaned English SRT, never from the raw auto track. Errors multiply in translation. Translate it yourself, pay a translator, or start with machine translation and have a native speaker check names, humor and idioms, the usual failure point.

In Studio, open Subtitles, choose Add language, and upload the translated file. Then add the translated title and description in the same section. That second step is the one people skip, and it matters as much as the captions. A perfect caption track behind an English title still gets skipped in a local feed.

What this unlocks is concrete. The same video can rank in a second language market with no reshoot, and viewers who would have left in 30 seconds stay. Pick languages from your data first, then invest in those one or two before adding more.

Keep expectations honest. A translated track will not fix a video built on local references nobody else gets, and machine translation alone will not sound native. But for how-to and travel content with clear speech, one good extra track is often the cheapest growth a small channel can buy.

Styling burned-in captions for a fast, easy read

Readability beats decoration. Viewers read captions in glances, often on a small screen in bright light. Every choice should serve that glance.

• Font size: On a 1080 by 1920 vertical export, start around 60 to 80 pixels and test on a real phone at arm’s length. If you squint, viewers already left.
• Font choice: Use a bold, plain sans-serif. Script fonts and thin weights look clever in the editor and fall apart at phone size.
• Contrast and box: Put text on a solid dark background box with a little padding. A shadow alone fails over bright footage. The box guarantees contrast on every frame.
• Words per line: Keep lines to roughly three to five words, one idea at a time. Long lines get cut off when a clip is cropped. Break at natural pauses.
• Color use: One main color for speech and, at most, one accent color for a key word or second speaker. Rainbow captions train viewers to ignore the emphasis.
• Placement: Keep captions in the lower middle or center, steady in one spot. Moving text forces viewers to hunt for it on every cut.

Most editors, including CapCut, Premiere Pro, Final Cut and DaVinci Resolve, can auto-transcribe speech into styled caption blocks. Use that as a draft, then apply the same cleanup pass as above. Save the settings as a preset so every clip matches, part of how viewers recognize your videos, as with a consistent opening when you are learning how to start vlogging.

Vertical safe zones: keeping text clear of app buttons

A 9:16 frame looks empty in your editor. It is not empty on a phone. TikTok, Reels and Shorts all cover the bottom with descriptions and titles, the right side with buttons, and the top with icons or search. Captions parked at the very bottom sit under all of that.

Keep key text inside the central area, roughly the middle 60 percent of the height, away from the right 15 percent where buttons stack. Most creators place captions just above or below the center, and preview the export inside each app. Many editors offer a safe zone overlay for Shorts, Reels and TikTok. Turn it on, and treat anything outside the lines as decoration only.

Plan for captions at filming time too. Leave headroom, keep the subject off the bottom of the frame, and keep the background behind the caption area plain. Our vertical video guide for Shorts, Reels and TikTok covers framing and export settings for these formats in full.

Transcripts: publish them, and let them work twice

Making one is easy once your SRT is clean. Remove the numbers and time codes, join the lines into sentences, add speaker names, and break the text into short paragraphs with a few plain headings. Do the paragraph pass yourself. A wall of unbroken text helps nobody.

For accessibility, a transcript serves people captions do not fully cover. A deaf-blind viewer using a screen reader needs text, not timed overlays. A viewer on a weak connection can read it when video will not load, and anyone can search and quote it in seconds.

For search, the transcript is your video in indexable form, holding the names, steps and phrases a viewer might type. Put it in a blog post that embeds the video, or link it from the description. Do not paste it onto several pages, and never publish a raw auto-caption dump.

Audio description basics for vloggers

Test your vlog with one question: could a listener who never sees the picture follow the story from audio alone? If you talk through what you do, name places as you arrive, and read out on-screen text, you are most of the way there. Gaps appear in montages, silent demos and reveals. Those gaps are where description helps.

You have three practical levels. The lightest is described narration while filming: say what you are showing as you show it. The middle option is a separate described-audio version, with extra narration in the pauses. The simplest fallback is a pinned comment or linked page with a full transcript plus notes on key visuals. It is not equal to real description, but it beats nothing.

Start with described narration as a habit, add transcripts for every video, and save full described-audio versions for your most visual uploads. That order keeps the work realistic for a small channel.

Caption questions, answered

Keep exploring

• How to Start a YouTube Channel
• How to Start Vlogging in 2026
• YouTube SEO Basics for Vloggers
• How to Edit Vlogs Faster
• Vertical Video Guide: Shorts, Reels and TikTok