Is my video uploaded to a server?

It depends on the mode. Browser mode runs Whisper entirely as WebAssembly on your device — the video never leaves it. Cloud mode sends the extracted audio (not the full video file) to a transcription endpoint for faster processing. Choose Browser mode if keeping the file off any server matters for your use case.

How do I get subtitles in Hindi, Punjabi, or another non-Latin script?

Select the spoken language explicitly from the dropdown before generating rather than leaving it on auto-detect. Whisper then transcribes in the correct native script — Devanagari for Hindi, Gurmukhi for Punjabi, Arabic script for Urdu — instead of a romanized guess that auto-detect sometimes produces for non-Latin languages.

How accurate is the transcription?

Whisper performs well on clear speech and common languages, but accuracy drops with heavy background noise, overlapping speakers, strong accents the model was undertrained on, or highly technical vocabulary and proper nouns. Always read through and correct the generated subtitles before burning them into a final video — treat the first pass as a strong draft, not a finished transcript.

Can I translate the subtitles into a different language?

Yes. After generating subtitles, use the Translate Subtitles panel to pick any target language. Translation runs through a free API and preserves the original timestamps, so translated subtitles stay in sync with the video without any manual re-timing.

What does "Burn & Download MP4" actually do to the file?

FFmpeg re-encodes the video and draws the subtitle text directly into the video frames as pixels, rather than storing them as a separate track. The resulting MP4 shows subtitles correctly on any player or platform, even ones that do not support separate subtitle files — but the trade-off is that the text can no longer be turned off, edited, or translated after burning.

How long does burning subtitles into a video take?

FFmpeg WASM runs single-threaded in the browser at roughly one to four times real-time speed depending on video resolution and device performance, so a one-minute 720p clip typically takes one to four minutes. Progress is shown throughout the re-encoding process.

Can I import subtitles I already have instead of generating new ones?

Yes. Use Import SRT to load an existing .srt or .vtt file, then translate it, apply one of the visual style presets, or burn it into the video — all without re-running transcription.

Is there a daily limit on how many videos I can subtitle?

AI transcription is metered per account — signed-in free accounts get a limited number of generations per day, and Pro accounts get a higher daily limit. Editing, restyling, and burning an already-generated subtitle set are not separately limited.

Why do the timestamps sometimes look slightly off after editing text?

Editing the text of a subtitle line does not change its timing — timestamps come directly from Whisper's speech recognition, not from the length of the text. If a line runs long relative to its display window, you can adjust it manually or shorten the wording so it reads comfortably within the time it is shown.

Does the tool work for videos that mix multiple languages?

Whisper handles some code-switching but performs best when a single primary language is selected for the whole video. For content that alternates significantly between languages, results are more reliable if you split the video into single-language segments before transcribing each one.

AI Subtitle Generator — Free Automatic Subtitles from Video (Browser)

FREE

Loading…

Go Pro ↑

Drop your video here

MP4, WebM, MOV, AVI — click to browse

Transcription Mode

Spoken Language

Always select the spoken language for accurate native-script subtitles (Hindi → Devanagari, Punjabi → Gurmukhi, etc.).
You can translate to any language after generation.

🎬

100% Private

Video never leaves your device

🌍

99+ Languages

Auto-detect or pick your language

🔥

Burn into MP4

Hard subs via FFmpeg WASM

Your files never leave your device. All processing happens locally in your browser — nothing is uploaded to our servers.Files are never stored on our servers · No tracking of your file contentsPrivacy Policy

What is AI Subtitle Generator?

Automatic subtitling works by running speech recognition over a video's audio track and producing timestamped text — one block of dialogue paired with a start and end time, repeated for the length of the video. This tool does that using OpenAI's Whisper model, in either of two modes: a cloud mode that sends the audio to a server for fast transcription, or a fully local mode that runs Whisper as WebAssembly directly in your browser tab so the video file never leaves your device. Both produce the same output format — an editable list of timed subtitle chunks you can correct, translate, restyle, and export.

Getting from raw transcription to usable subtitles involves more than just recognizing words. Timing matters: a subtitle that lags behind speech or lingers too long after it ends is distracting in a way that is hard to unsee once you notice it, so the chunk boundaries here are generated directly from Whisper's word-level timestamps rather than guessed from average reading speed. Script and language matter too — selecting the spoken language explicitly (rather than leaving it on auto-detect) makes the difference between getting subtitles in the correct native script for Hindi, Punjabi, Arabic, or Japanese versus getting a romanized approximation that native speakers find awkward to read.

Once the raw transcript exists, you can translate it into a different target language (useful for reaching an audience that does not speak the original language), restyle it with one of six visual presets, and either export a standard SRT/VTT subtitle file for use in a video editor or YouTube, or have FFmpeg (also running as WebAssembly) permanently burn the subtitles into a new MP4 so they display correctly on any player without needing a separate subtitle file at all.

How It Works

Upload a video and pick a transcription mode. Cloud mode uploads just the audio to a transcription endpoint and returns timestamped segments in seconds — the faster option for longer videos or slower devices. Browser mode extracts and resamples the audio locally, then runs it through a Whisper WebAssembly worker entirely on your machine, which is slower but keeps the file completely off any server. Either way, the output is a list of timed text chunks you can click through and edit individually, with a live overlay on the video preview showing exactly how each line will look. From there, optionally translate the whole subtitle set to a different language using a free translation API, pick one of six font presets (Classic, Bold, Neon, Cinema, Handwritten, Retro) and a screen position, then either export SRT or VTT files for use elsewhere, or click "Burn & Download MP4" to have FFmpeg re-encode the video with the subtitles drawn permanently into the video frames.

How to Use

Upload your video

Drop an MP4, WebM, or MOV file onto the upload area, or click to browse. There is no file size limit enforced by the tool itself, though very long videos take proportionally longer to transcribe.

Select the spoken language explicitly

Choose the language actually spoken in the video rather than leaving it on auto-detect. This matters most for non-Latin scripts — setting Hindi, Punjabi, Arabic, or similar explicitly ensures subtitles come back in the correct native script instead of a romanized guess.

Generate the transcript

Click Generate. Cloud mode returns a transcript in seconds by processing audio on a server; Browser mode runs Whisper locally in WebAssembly, which keeps the video entirely private but takes longer, especially on older hardware.

Edit and correct the subtitles

Click any line in the list to fix transcription errors, awkward line breaks, or punctuation. Click a timestamp to jump the video preview to that exact moment and check the wording against what is actually said.

Translate if you need a different output language

Use the Translate Subtitles panel to convert the whole set into another language while keeping the original timestamps intact — useful for repurposing a video for a different-language audience without re-transcribing from scratch.

Pick a font style and screen position

Choose from six visual presets and a Top, Middle, or Bottom position. The live overlay on the video preview shows exactly how burned-in subtitles will look before you commit to rendering.

Export SRT/VTT, or burn subtitles into a new MP4

Download a standard subtitle file for use in YouTube or a video editor, or click Burn & Download MP4 to have FFmpeg WASM permanently draw the subtitles into a new video file that displays correctly everywhere without a separate subtitle track.

Who Uses This Tool

🎬

Content Creators & YouTubers

Add burned-in captions to short-form video for platforms where autoplay defaults to muted, since a large share of viewers only watch with sound off.

🌍

Reaching a Non-Native-Language Audience

Transcribe in the original spoken language, then translate the subtitle set to reach viewers who do not speak it, without re-recording anything.

📚

Students & Researchers

Turn recorded lectures or interviews into a searchable, timestamped transcript that is far faster to review and quote from than re-watching the full recording.

♿

Accessibility

Provide captions for deaf and hard-of-hearing viewers, which is both a genuine accessibility improvement and, on many platforms, a ranking or compliance factor.

🏢

Training & Internal Video

Caption onboarding or training videos so multilingual teams and non-native speakers can follow along accurately without relying on live audio comprehension.

🎙️

Podcasters Repurposing Audio to Video

Turn a podcast episode into a captioned video clip for social platforms, where captions are frequently the reason someone stops scrolling and watches.

Why Choose This Tool

A Genuine Choice Between Cloud Speed and Local Privacy

Cloud mode is faster for long videos; Browser mode keeps the file entirely on your device by running Whisper as WebAssembly. You are not locked into one trade-off.

Correct Native Script for 99+ Languages

Setting the spoken language explicitly returns transcripts in the correct script — Devanagari, Gurmukhi, Arabic, CJK — rather than a romanized approximation from auto-detect.

Free Translation Without Leaving the Tool

Translate a finished subtitle set into a different target language while keeping the original timestamps, instead of exporting, translating elsewhere, and re-importing.

Burn Subtitles In, or Keep Them as a Separate File

Export a standard SRT/VTT file for editors and platforms that support subtitle tracks, or permanently burn captions into the video for platforms and players that do not.

Editable Before You Commit

Every subtitle line is editable and clickable-to-seek before export, so transcription mistakes get caught and fixed before they end up baked into a final video.

Six Visual Styles Built for Different Content Types

From a clean Classic look to a bold YouTube-style caption or a cursive Handwritten font, the preset styles cover the range most creators actually reach for.

Key Benefits

Whisper-based transcription supporting 99+ languages with correct native script output

Choice of cloud transcription (fast) or fully local browser transcription (private)

Word-level timestamp accuracy carried through to subtitle chunk boundaries

Free subtitle translation into any target language after generation

Six font presets — Classic, Bold, Neon, Cinema, Handwritten, Retro — plus position control

Live subtitle overlay preview synced to video playback

Interactive timeline for clicking directly to any subtitle's moment in the video

Edit any subtitle line before exporting or burning

Export standard SRT/VTT, or burn subtitles permanently into a new MP4 via FFmpeg WASM

Import an existing SRT file to restyle or re-burn without re-transcribing

Frequently Asked Questions

Free Online Document Tools for Everyone

iLoveDocsTools is used by students, professionals, and businesses across the United States, Canada, United Kingdom, Australia, India, and more than 100 other countries. All tools are browser-based and process files privately — your files never leave your device. Free to use, with Pro plans for unlimited access.

AI Subtitle Generator

What is AI Subtitle Generator?

How It Works

How to Use

Who Uses This Tool

Why Choose This Tool

Key Benefits

Frequently Asked Questions

Related Tools

Free Online Document Tools for Everyone

No Upload Required

Free to Use — Pro Plans Available

Works on Any Device