SRT Generator — Create SRT Subtitle Files from Video or Audio (Free)
Upload a video (MP4, MOV, WEBM) or audio file (MP3, WAV, M4A). VexaScribe transcribes with word-level timestamps and gives you a standards-compliant SRT with HH:MM:SS,mmm timecodes and UTF-8 encoding — ready for YouTube, Premiere Pro, DaVinci Resolve, or VLC. 30 minutes free, no card.
Supported formats:
By VexaScribe Editorial · Verified
TL;DR
VexaScribe generates SRT subtitle files from video or audio in three steps: upload, preview the auto-split cues, download. The output is a plain-text .srt with numbered blocks, HH:MM:SS,mmm timecodes (comma before milliseconds), and UTF-8 encoding — the exact SubRip format YouTube, Premiere Pro, DaVinci Resolve, and VLC expect. Free for the first 30 minutes.
How to Create an SRT File
Upload video or audio
Drop MP4, MOV, WEBM, MP3, WAV, M4A, FLAC, OGG, or OPUS. Files up to 5GB supported. Paste-URL flow also works.
Preview the SRT cues
VexaScribe transcribes with word-level timestamps and splits into cues at ~80 characters and 5 seconds. Fix words or drag timings in the editor.
Download as .srt
Standards-compliant SRT with HH:MM:SS,mmm timecodes and UTF-8 encoding. No watermarks. Ready for YouTube, Premiere, DaVinci, VLC.
What You Get
- • Standards-compliant SRT — numbered cue blocks,
HH:MM:SS,mmm --> HH:MM:SS,mmmtiming lines (comma before milliseconds, per the SubRip spec), one blank line between cues. - • UTF-8 encoding, no BOM — accented characters, non-Latin scripts (Cyrillic, Arabic, Chinese, Japanese) render correctly across players.
- • Word-level cue splitting — cues cap at ~80 characters and 5 seconds (10s hard ceiling) so files import cleanly into YouTube, Premiere, DaVinci, and VLC without manual cleanup.
- • Live preview — see and edit the cues before you download.
- • No watermarks, no attribution requirement, commercial use allowed on the free tier.
Sources You Can Generate SRT From
From video
MP4, MOV, WEBM, MKV, AVI. VexaScribe extracts the audio track automatically before transcribing. If you need format-specific notes (H.264 audio channels, MOV timecode drift, WEBM Opus quirks) see the video to SRT guide.
From audio
MP3, WAV, M4A, FLAC, OGG, OPUS. Podcasts, voice memos, meeting recordings, interviews. For format compatibility, accuracy by scenario, and podcast-specific workflows, use the dedicated audio to SRT page. Generating SRT for an entire podcast season or webinar series? Use bulk transcription — up to 50 files per batch, all exported as SRT in one download.
From a YouTube URL
Paste any public YouTube URL and export the captions (or a fresh transcription) as SRT. Full workflow at YouTube subtitle download.
From text (with timing)
Have a script but no audio? SRT requires per-cue timing, which text alone doesn't provide. Two options: record a reference read of the script and let the generator align cues to your voice, or hand-type timecodes in a text editor following the format at what is an SRT file.
SRT File Format — Quick Reference
An SRT file is plain text. Each cue is four parts: a sequence number, a timing line with --> between start and end, one or two lines of caption text, and a blank line before the next cue.
1
00:00:01,000 --> 00:00:04,500
Welcome to today's presentation.
2
00:00:04,800 --> 00:00:08,200
We'll be covering three main topics.Full spec (encoding, styling, line-length conventions, developer parsing notes, SRT vs VTT vs ASS comparison table): what is an SRT file.
Editing the SRT After Generation
For quick text fixes, any editor works — Notepad on Windows, TextEdit (in plain text mode) on Mac. For anything more, VS Code is the recommended choice because it shows the file encoding in the status bar and lets you re-save as UTF-8 without BOM in one click — the most common cause of garbled characters in shipped SRTs.
- • Encoding fixes — VS Code: click the encoding indicator in the bottom-right → “Save with Encoding” → UTF-8.
- • Waveform-based retiming — Subtitle Edit (Windows, free) or Aegisub (cross-platform, free). Both align cues to a waveform view.
- • Line length — the broadcast convention is 42 characters per line, max 2 lines per cue (Netflix Timed Text Style Guide, BBC Subtitle Guidelines). Longer lines are technically valid but hard to read at caption font size.
- • Timecode format — must be
HH:MM:SS,mmmwith a comma before milliseconds. A period is WebVTT, not SRT — some players silently fail on the wrong separator.
Need the SRT in another language? Use the dedicated subtitle translator — timestamps stay byte-identical to the source, formatting tags (italics, positioning) preserved through translation, 133 target languages. Different from generating an SRT from scratch (which is this page's job).
Using Your SRT File
Premiere Pro
File → Import → select .srt. Cues land on a caption track — style, reposition, and burn in from there.
DaVinci Resolve
File → Import Subtitle. Resolve creates a subtitle track you can edit and colour-grade with the timeline.
Final Cut Pro
File → Import → Captions. Adjust styling and timing within FCP's caption editor.
YouTube
YouTube Studio → Content → your video → Subtitles → Add Language → Upload File → With timing → select .srt.
Vimeo
Add SRT captions in Vimeo's Advanced settings during publish. Multi-language toggle supported.
VLC / MPV
Name the .srt to match the video (video.mp4 + video.srt) and the player auto-loads it on playback.
Facebook / LinkedIn
Attach the SRT during publish. Captions auto-display for muted feed playback.
LMS platforms
Teachable, Thinkific, Kajabi, Canvas, Moodle — all accept SRT for course video captions.
SRT Accuracy — What to Expect
VexaScribe runs Whisper Large-v3 (OpenAI, MIT license). The original paper (Radford et al., arXiv:2212.04356) reports ~4.2% WER on LibriSpeech clean — a lab benchmark of well-recorded audiobook narration. Real-world figures we measure on user files:
- • 92–95% — clean single-speaker English recorded with a proper microphone (Shure SM7B, Rode PodMic, Blue Yeti).
- • 85–92% — two-speaker interview, both parties mic'd.
- • 75–85% — accented English (non-native, regional).
- • 70–82% — technical, medical, or legal vocabulary.
- • 65–78% — noisy environments (café, street, echoic room).
Overlapping speakers, phone-line audio (8 kHz sampling), and strong accents are the main drivers of error. Review the transcript before publishing for anything customer-facing.
When to Hire a Human
Broadcast delivery, legal depositions, medical records, courtroom transcripts, and accessibility compliance for regulated industries need human review or human authorship. Whisper Large-v3 gets you 95% of the way there quickly; the last 5% is where mishearings on names, dosages, jurisdictions, and citations create legal risk. For a captioned marketing video, a lecture recording, a podcast episode, or a YouTube upload, auto-generated SRT with a quick review pass is the right tool.
Affordable Pricing
Why Use VexaScribe for SRT Generation?
Everything you need to create professional subtitle files.
Whisper Large-v3 accuracy
Powered by Whisper Large-v3 (Radford et al., arXiv:2212.04356) — ~4.2% WER on LibriSpeech clean, 92–95% on real-world clean English podcasts.
Word-level cue timing
Cue boundaries land on real word starts and ends — not linear interpolation. Pauses stay on screen instead of becoming sub-second flashes.
In-browser SRT editor
Edit text, drag start and end times, split or merge cues, preview timing — no software to install.
Speaker labels in cues
Speaker diarization prefixes cue text with SPEAKER 1, SPEAKER 2, etc. Renamable before export.
Batch processing
Upload multiple files in one session and export SRT for all of them. Useful for content libraries.
SRT, VTT, TXT, DOCX
Same transcription, four export formats. Convert to VTT for HTML5 web players in one click.
Frequently Asked Questions
How do I create an SRT file?
Three steps. (1) Upload a video (MP4, MOV, WEBM) or audio (MP3, WAV, M4A) file. (2) VexaScribe transcribes speech with millisecond-level word timestamps. (3) Preview the cues, correct any words, then click Download as SRT. Total wall time is roughly the length of your file divided by five — a 10-minute clip is ready in about two minutes. The output is a standards-compliant .srt with HH:MM:SS,mmm timecodes and UTF-8 encoding.
What is the SRT file format?
SRT (SubRip Subtitle) is a plain-text subtitle file. Each cue is a numbered block containing a sequence integer, a start-and-end timecode line in HH:MM:SS,mmm format with a comma before milliseconds (unlike WebVTT which uses a period), and one or two lines of caption text. Cues are separated by a blank line. Full spec, working code sample, and encoding notes at /what-is-an-srt-file.
Can I create an SRT file from audio?
Yes. Upload MP3, WAV, M4A, FLAC, OGG, or OPUS — the generator transcribes the speech and emits SRT with per-cue timestamps. If your input is audio-only (podcast, voice memo, meeting recording), the dedicated /audio-to-srt page walks the format-specific quirks in more detail.
Can I create an SRT file for free?
Yes, up to 30 minutes on signup with no card required. Paid plans start at $2/month for 200 minutes; higher tiers scale to 6,000 minutes/month at $20. Whisper Large-v3 (which powers the transcription) is MIT-licensed and free to self-host if you have Python and a GPU, but the DIY route is slower and lacks the cue-splitting and editor.
What accuracy can I expect?
About 92–95% on clean single-speaker English recorded with a proper microphone, dropping to 75–85% on accented English and 65–78% on noisy environments. These numbers reflect Whisper Large-v3 real-world performance; the underlying paper (Radford et al., OpenAI, arXiv:2212.04356) reports ~4.2% WER on the LibriSpeech clean benchmark. Review the transcript before publishing, especially for technical vocabulary or overlapping speakers.
How do I edit an SRT file after generating?
You can fix cues in the VexaScribe editor before download — change text, drag timestamps, split or merge cues. After download, any text editor works (Notepad, TextEdit, VS Code). For encoding fixes, VS Code is recommended: open the file, click the encoding indicator in the status bar, and save as UTF-8 without BOM. For visual timing work (waveform alignment), use Subtitle Edit (Windows, free) or Aegisub (cross-platform, free).
Does this work for Premiere Pro, DaVinci Resolve, or iMovie?
Yes. Premiere Pro imports SRT via File → Import — the cues land on a caption track. DaVinci Resolve uses File → Import Subtitle. Final Cut Pro imports via File → Import → Captions. iMovie does not import SRT directly; the workaround is to burn the captions into the video first (Descript, HandBrake with subtitle burn-in, or FFmpeg).
Do I keep the file if I use the free trial?
Yes. The SRT you generate during the 30-minute free tier is yours to download and use commercially. No watermark, no attribution requirement. If you upgrade later, previously generated files stay in your workspace and can be re-exported.
Note: Transcription accuracy depends on audio quality, background noise, and speaker clarity. Review generated SRT files before publishing.
Related Tools
Audio to SRT
Convert MP3, WAV, M4A, FLAC, OGG audio to SRT — format compatibility, accuracy by scenario, podcast use cases.
Video to SRT
Generate SRT from MP4/MOV/WEBM video, or add an existing SRT to a video (soft-subs vs burn-in + FFmpeg).
What Is an SRT File?
Format spec — working code sample, HH:MM:SS,mmm timecodes, UTF-8 encoding, SRT vs VTT vs ASS reference.
YouTube Subtitle Download
Pull captions from any public YouTube video as SRT, VTT, or TXT.
Subtitle Generator
Generate SRT and VTT — plus optional burn-in for platforms that don't accept caption files.
Subtitle Translator
Already have an SRT? Translate to 133 languages, timestamps preserved. Batch up to 50 files.
Video to Text
Video → plain-text transcript (no timing) with speaker labels.
How to Generate Subtitles from Video
Complete 4-step workflow: extract, transcribe, export SRT/VTT, attach.
Add SRT to MP4
You have the SRT — now embed it. FFmpeg, HandBrake, VLC, Premiere Pro, and 5 online tools compared.