How to Create Subtitles for Any Video (SRT/VTT) in 2026

Your video is done. Now it needs subtitles — and you'd rather not learn Premiere Pro or pay for yet another monthly subscription just to get one SRT file. Good news: you don't have to. Paste a link or upload the file to an AI transcription tool, wait a couple of minutes, export SRT or VTT. That's the whole workflow. Below: what those formats actually are, how to attach them to YouTube or TikTok, and what to double-check before publishing.

What is an SRT file and how does it work?

An SRT file (SubRip Subtitle) is a plain text file that pairs numbered blocks of text with timecodes. It tells a video player exactly what to display and when. Each block — called a cue — has three parts: a sequence number, a start and end timestamp, and one or two lines of text.

Here's what a cue looks like inside the file:

1
00:00:03,120 --> 00:00:06,540
Welcome back to the channel.
Today we're talking about captions.

That's it. No fonts, no colors, no video data — just text and timing. This simplicity is why SRT became the universal standard: YouTube, LinkedIn, Vimeo, Premiere Pro, DaVinci Resolve, CapCut, and VLC all read it. You can even open an SRT in Notepad and fix a typo by hand.

SRT vs VTT: which subtitle format do you need?

VTT (WebVTT) is SRT's younger web-native sibling. The cue structure is nearly identical — the main visible differences are the WEBVTT header at the top and periods instead of commas in timestamps. The practical differences matter more:

The rule of thumb: uploading to a platform or importing into an editor → SRT; embedding video on your own website → VTT. When in doubt, generate both — a good tool exports either format from the same transcript, along with TXT and DOCX if you also need the text for a blog post or show notes.

How do you generate subtitles automatically? (Step-by-step)

Here's the fastest route from "video with no captions" to "ready SRT file" using VideoScribe:

1. Paste a link or upload a file. YouTube, Instagram, and TikTok URLs work directly — no downloading required. Local files up to 500 MB can be uploaded straight from your device.
2. Wait 1–3 minutes. The AI engine transcribes the audio and returns a full transcript with timestamps. No registration needed for the free tier.
3. Export SRT or VTT. One click, and the file is ready to upload to YouTube Studio or drop into your editor.

A note on what's free and what's not. Transcription itself is free — 3 videos per day, up to 30 minutes each, PDF export included. SRT and VTT export is a paid feature: a one-time credit from 29 UAH (~$0.70) covers a video up to 3 hours. No subscription — you buy a credit, use it, done. Bundles cost less per video: 5 for 99 UAH (~$2.40) or 10 for 179 UAH (~$4.30). Full details are on the pricing page.

The manual alternative: typing subtitles yourself

You can also create subtitles by hand in YouTube Studio's caption editor or a free tool like Aegisub. Pros: it costs nothing, and you control every word and line break. Cons: it's brutally slow. One minute of video takes 5–10 minutes of manual work, so a 20-minute video eats half your afternoon. Hand-timing makes sense for a 60-second ad where every frame counts. For anything longer, AI gets you 95%+ of the way there in two minutes — you proofread instead of typing.

How do you add an SRT file to YouTube, TikTok or a video editor?

Once you have the file, attaching it takes a minute:

• YouTube: YouTube Studio → Subtitles → select your video → Add Language → Upload file → "With timing" → choose your SRT. Done — viewers can toggle captions with the CC button.
• CapCut / InShot: import the SRT via the Captions or Text menu (CapCut desktop: Text → Local captions → Import). The cues appear as editable text layers you can restyle.
• Premiere Pro / DaVinci Resolve: File → Import (Premiere) or File → Import → Subtitle (Resolve). Both let you burn the captions into the video or export them as a sidecar track.
• Your own website: use a VTT file with the HTML5 <track> tag:

<video controls src="lesson.mp4">
  <track kind="captions" src="lesson.vtt" srclang="en" label="English" default>
</video>

TikTok's mobile app doesn't accept SRT uploads directly, so for TikTok the usual move is burning subtitles in via CapCut before posting.

Not sure the flow fits your video? The easiest check is to run one through VideoScribe on the free tier — you'll see the transcript quality before spending anything.

How accurate are AI-generated subtitles?

On clean audio — one speaker, decent mic, minimal background noise — modern AI transcription reaches 95–99% accuracy. VideoScribe runs on AssemblyAI with a Whisper fallback, supports 98+ languages, and can label individual speakers, which is a lifesaver for interviews and podcasts.

That last 1–5% is where you should spend your proofreading time. The usual suspects:

• Proper nouns: names of people, brands, and places get creative spellings.
• Industry jargon and acronyms: "SRT" itself might come out as "S-R-T" or "search."
• Numbers and units: "2026" vs "twenty twenty-six."
• Crosstalk: overlapping speakers reduce accuracy on those segments.

A realistic expectation: for a 15-minute talking-head video, plan on 3–5 minutes of proofreading. Still an order of magnitude faster than typing from scratch.

What makes good subtitles? Formatting rules

Accurate text with bad formatting still reads poorly on screen. The conventions professional subtitlers follow:

• Max 42 characters per line, two lines per cue. Longer lines force the eye to scan instead of glance.
• Keep each cue on screen at least 1.5 seconds, even for short phrases — viewers need time to register the text.
• Break lines at natural pauses, not mid-phrase. "He opened the door / and walked out" beats "He opened the / door and walked out."
• Don't split a person's name or a number across lines.

Why bother? Because much of social video is watched with the sound off — muted feeds are the default on phones. Captions are what keep those viewers from scrolling past. Captioned videos hold attention longer, and subtitles make your content accessible to deaf and hard-of-hearing viewers — and searchable by text.

Create your first subtitle file in 2 minutes

The full pipeline — paste a link, get a timestamped transcript, export SRT or VTT, upload to your platform — genuinely takes a couple of minutes plus a quick proofread. Try it free right now: VideoScribe gives you 3 transcriptions a day at no cost. When you need the actual subtitle file, a one-time credit costs less than a dollar. No subscription, no account required to start.

FAQ

How do I create an SRT file for free?

Transcription on VideoScribe is free — 3 videos per day up to 30 minutes each, with free PDF export of the transcript. Exporting the SRT or VTT file itself is a paid feature, but it's a one-time credit from 29 UAH (~$0.70) per video, not a subscription.

What is the difference between SRT and VTT?

VTT is built for the web: HTML5 players read it natively via the <track> tag, and it supports styling like text color and positioning. SRT has no styling but is accepted almost everywhere — YouTube, video editors, media players. Use SRT for platforms and editors, VTT for your own website.

Can YouTube generate subtitles automatically?

Yes, but auto-captions often miss punctuation, mangle names and technical terms, and aren't available in every language. Uploading your own corrected SRT gives you accurate, properly punctuated captions — and you keep a reusable file for other platforms.

How long does it take to generate subtitles?

Typically 1–3 minutes for most videos. You can paste a YouTube, Instagram, or TikTok link or upload a file up to 500 MB; paid credits handle long videos up to 3 hours (or up to 8 hours with a long-video credit for 79 UAH).