What is an SRT file?

What is inside an SRT file

SRT stands for SubRip Text, named after the late-1990s Windows program that ripped subtitles off DVDs with OCR. The program is a historical footnote; the format it wrote conquered the world. There is no official standards document — unlike WebVTT (a W3C standard), SRT is a de facto standard that won by simplicity: a human can fix a typo in Notepad, a script can generate one in three lines of Python, and every player, editor, and upload form accepts it.

Each cue block has exactly four parts:

Two properties define the format's character: it is tiny (a feature film's subtitles are a few dozen kilobytes of text) and it is encoding-agnostic — the format sets no character encoding, so modern files are written UTF-8 while DVD-era files arrive in Windows-1252 or regional encodings, which is where the famous mojibake (garbled accented characters) comes from. There is no header line (that is VTT's WEBVTT), no metadata, and no frame-rate declaration — millisecond precision means you convert frames to milliseconds yourself if a tool needs that.

What opens an SRT file

The answer depends on the job: read and fix the text, watch the subtitles on the video, or edit timing on a waveform. Any editor opens an SRT instantly; a player shows it timed with the video; a subtitle editor does both.

The double-click trap: Windows has no built-in SRT association — double-clicking a downloaded subtitle file opens an app-picker or nothing, because the OS does not know which tool claims a subtitle file. It is not a video (players will not grab it) and not a document (word processors open it with formatting damage). Right-click → Open with → Notepad for text, or drop it on a playing VLC window for the timed view.

VLC specifics (the most-asked variant of this question): put the SRT in the same folder as the video with the same base name — Movie.mp4 next to Movie.srt — and VLC auto-loads it. If nothing shows, press V to cycle subtitle tracks (VLC remembers if you disabled tracks last session), or use Subtitle → Add Subtitle File. Boxes and question marks instead of characters mean the encoding setting: Tools → Preferences → Subtitles/OSD → Default encoding, matched to the file (UTF-8 for anything modern). Drift is Tools → Track Synchronization, in milliseconds.

iPhone and Android: mobile players are the weakest SRT surface — iOS's built-in video playback makes attaching external subtitle files awkward (VLC's iOS app is the practical route), while Android's MX Player and VLC auto-load an SRT that sits beside the video with a matching name, same as desktop. On phones, a downloaded subtitle's encoding more often shows as boxes — fix it on a desktop first.

A sample SRT file you can copy

This is a complete, valid SRT file — two cues, every rule applied. Save it as sample.srt and drop it next to any video renamed sample.mp4 to see it work in VLC, or upload it to YouTube's caption editor as a first test:

1
00:00:01,000 --> 00:00:04,500
Welcome to the sample subtitle file.

2
00:00:05,200 --> 00:00:09,800
Each block is a sequence number,
a timecode line, then the text.

For a large realistic file (a 45-minute talk is roughly 500–800 cues and 40–70 KB), public sample repositories offer test files in several sizes — SampleFile.com hosts SHA256-verified SRT samples from 50KB to 1MB with a stable manifest for test pipelines, ConvertICO offers small themed samples (dialogue, lyrics, multilingual unicode), and the andreyvit/subtitle-tools GitHub repo carries the classic two-cue sample.srt that tools use in their own docs. These matter more than they look: a known good file is the fastest way to prove the player is the problem, not your subtitle file.

Generating a file for real video? That is the how-to-create-an-SRT-file guide — hand-writing, waveform timing, and generating from audio, with the Notepad and TextEdit save traps.

When an SRT file breaks (and the fix)

SRT fails silently — the parser skips or mis-renders cues without an error. The recurring break modes and their fixes:

The SRT vs VTT guide covers the format-decision side of these traps — where VTT is required instead (any HTML5 <track> element), and the one-line ffmpeg conversions between the two.

Turning an SRT file into plain text

The reverse job — you have an SRT and want the transcript — is a text filter: keep the caption lines, drop the numbers and timecodes. Browser-local converters do it in one paste (SubtitleKit keeps or strips timecodes, files never leave your device); for scripts, Python's srt library (pip install srt) parses cue objects you can join into text, and one-liner gists handle the common “strip timestamps into one paragraph” case in ten lines.

VocalFuse inverts the pipeline: instead of extracting text from subtitles that already exist, it runs a local Whisper-class engine on the audio itself — no captions required — and exports SRT, VTT, or TXT from one pass. When the source video never had captions (the ~1-in-10 case), there is no SRT to open; the transcript has to come from the audio. Nothing uploads, and there is no per-minute meter: free tier, Basic $5/mo dictation, Pro $10/mo adds AI notes and summaries.

Related reading

SRT file — FAQ