Direct answer: You can transcribe video privately: your video file stays on your device while your browser extracts compressed audio and sends it to a transcription provider. The audio is deleted right after processing and never stored. On-device transcription is an optional mode on capable devices, never downloaded by default. Edit the transcript, then export TXT, SRT, VTT, or MP4, free with no signup.
OneDrop Video Captions is a free browser-based AI video caption generator that transcribes video while your video file stays on your device. (It is not related to One Drop, the diabetes-health company. Similar-sounding name, completely different product.)
Transcription is the step that turns speech into text with timestamps. Almost every transcription tool starts by uploading your whole video file to remote servers. OneDrop Video Captions works differently: your video file never leaves your device. Your browser extracts compressed audio, sends it to a transcription provider, and the audio is deleted right after processing, never stored. You still get a clean, editable transcript. This page explains how it works, what your device needs, how accurate it is, and how to export the result.
How does browser-based transcription work?
Browser-based transcription keeps your video file on your device and uses a speech-recognition AI model, Whisper, via a transcription provider instead of uploading your footage. Here is what happens under the hood:
- Your browser extracts the audio. When you drop in a video, the browser decodes it in your browser and pulls out the audio track. No video file is uploaded.
-
The provider transcribes the audio. Compressed audio goes to a transcription provider, which listens in small chunks and writes out what it hears, with timing information for every word and automatic punctuation. The audio is deleted right after processing and never stored.
-
Uncertain words get flagged. The transcript marks words it has low confidence in, so you know exactly where to double-check instead of re-reading the whole transcript.
-
You edit the transcript. Every word is clickable. Fix names, terms, and mistakes, and the timing stays attached to the corrected text. You need an internet connection while transcribing, because the transcription provider does the AI work. What never happens: your video file is not uploaded, copied to a server, or retained anywhere.
What do you need to transcribe video in your browser?
The requirements are modest:
-
A modern browser. Any recent version of the major browsers on desktop or mobile can run the studio. If your browser can play the video file, it can usually transcribe it.
-
An internet connection while transcribing. The transcription provider does the AI work, so you need to be online. Your video file still never leaves your device.
-
A supported video file. MP4, MOV, or WebM, up to 2 minutes per video on the free plan. Paid plans raise the limit to 30 minutes per video. You do not need to install anything or create an account. If a tool demands your whole video file uploaded before transcribing, it is not keeping your file on your device.
How accurate is browser-based transcription, and how do you get better results?
For clear, close-microphone speech, Whisper transcription via a provider handles punctuation and word timing well out of the box. Like every speech-recognition system, it has weak spots:
-
Proper names and brand terms. Uncommon names are the most frequent mistakes. They are also the easiest to fix, since you know what they should say.
-
Slang and heavy accents. Casual speech and strong accents produce more errors than careful, neutral speech.
-
Noisy recordings. Background music, wind, and room echo all make the model's job harder. The studio detects where speech actually happens and leans on the voice rather than the music. If music truly covers the voice, it tells you that instead of inventing words.
-
Overlapping speakers. Two people talking at once is hard for every transcription engine. How to improve accuracy before you even press transcribe:
-
Record close to the microphone. This single change improves results more than anything else.
-
Reduce background noise. A quiet room beats any software cleanup.
-
Speak at a natural, steady pace. Rushed or mumbling speech costs accuracy.
-
Pick the right language first. You choose English, Hindi, and Urdu (including Hinglish and Roman Urdu) before transcribing, including mixed Hindi-English and Urdu-English speech (Hinglish and Roman Urdu), which comes back in English, Hindi, and Urdu (including Hinglish and Roman Urdu) without being translated.
-
Do a read-through edit. The low-confidence flags point you to the exact words worth checking. A two-minute review catches the stragglers. If you need maximum accuracy on difficult audio, Pro ($13/mo) and Studio ($20/mo) include a "Boost accuracy" mode (Whisper large-v3 on our servers): an explicit opt-in capped at 30 min/day on Pro and 120 min/day on Studio. Boost uses your monthly caption minutes.
From transcript to finished subtitles: exporting TXT, SRT, VTT, and MP4
A transcript becomes useful when it leaves the tool in the right format. OneDrop Video Captions exports four formats from the same project:
-
TXT: a plain-text transcript. Best for blog posts, show notes, documentation, and searching the content of a video.
-
SRT: the standard subtitle file for YouTube, Vimeo, LinkedIn, and most video players. Carries text plus start and end times.
-
VTT (WebVTT): the web-native subtitle format, preferred by HTML5 players and many learning platforms. Supports basic styling cues that SRT does not.
-
MP4: your video with the captions burned directly into the frames, in 9:16 or 16:9. Best for TikTok, Reels, and Shorts, where uploaders cannot rely on the platform to display sidecar subtitles. Exporting takes four steps:
-
Finish your read-through edit so names and terms are correct.
-
Choose your format based on where the video is going (MP4 for social feeds, SRT or VTT for platforms that attach subtitles themselves, TXT for text uses).
-
For SRT and VTT, skim the timing on a few captions. AI timing is usually frame-close, but long pauses sometimes need a nudge.
-
Download. The files generate in your browser on your device with no watermark. Not sure which subtitle format to pick? Our SRT vs VTT guide walks through the differences.
Privacy: what "your file stays on your device" actually covers
This is a specific technical claim, so here is exactly what it covers in the free OneDrop Video Captions studio:
-
Your video file is decoded in your browser tab. It is not transmitted to a server.
-
Your browser extracts compressed audio and sends it to a transcription provider. The audio is deleted right after processing and never stored.
-
Caption rendering and MP4 export happen in your browser. The finished files are generated on your device.
-
No account is required, so there is no user profile for your projects to attach to.
-
On-device transcription is an optional mode on capable devices, never downloaded by default and never required. What it does not cover: the paid features. Saved projects and cloud storage on paid plans require an account and do store data on servers. The server-side "Boost accuracy" mode (Pro and Studio) sends audio to a server by design. These are paid, clearly labeled, and separate from the free workflow.
A final honest note: browser-based processing keeps your footage off other companies' servers, which removes the biggest privacy risk in captioning. It does not make claims about your device's own security, and no tool should promise that. Judge every caption tool by what it verifiably does with your files, not by adjectives on its homepage.
FAQ
Is there an offline mode?
On-device transcription is an optional mode on capable devices, never downloaded by default and never required. The standard workflow needs an internet connection to reach the transcription provider.
Which languages can I transcribe?
English, Hindi, and Urdu, including mixed Hindi-English and Urdu-English speech (Hinglish and Roman Urdu). Hindi and Urdu come back in their own scripts or as Roman Hindi/Roman Urdu typed the way phones spell it, without translation.
How long a video can I transcribe for free?
Up to 2 minutes per video, within 2 minutes of audio per day, on the free plan. Paid plans raise the per-video limit to 30 minutes and add more monthly minutes.
Can I edit the transcript?
Yes. Every word is clickable. You can correct text, fix timing, mark emphasis, censor profanity, and restore silent gaps before exporting.
What is the difference between a transcript and subtitles?
A transcript is the full text of what was said, usually as a TXT file. Subtitles are that text split into timed captions, as SRT or VTT files, so video players can display each line at the right moment. The studio produces both from the same transcription.
Is my audio sent anywhere during transcription?
In the free studio, your video file never leaves your device: the browser extracts compressed audio and sends it to a transcription provider, and the audio is deleted right after processing. The "Boost accuracy" mode on Pro and Studio is an explicit opt-in that also uses server-side transcription and is clearly labeled as such.
Transcribe your first video free
Drop in a video and get an editable transcript in your browser. No video-file upload, no signup, no watermark.