Questions, answered plainly

OneDrop is a browser-based captioning studio. You drop in a video, it transcribes the speech, the video file never leaves your device; only compressed audio is sent to the transcription provider and deleted right after and you edit the text, pick a caption style and export an MP4 with burned-in captions or an SRT, VTT or TXT file.

No. Open the studio, drop a video and start captioning, no signup, no email, no trial. Accounts arrive with the paid plans for saved projects and cloud storage, but the free studio will stay account-free.

Any modern browser with WebAssembly, current versions of Chrome, Edge, Firefox and Safari all work. Desktop is recommended for long videos; shorter videos caption fine on recent phones.

No. The video file never leaves your device, it is decoded and the captions rendered inside your browser tab; only compressed audio is sent to a transcription provider and is deleted right after processing. On-device processing is available as an opt-in on capable devices.

The studio keeps your recent projects, style choices and edits in your browser's local storage so you can come back to them. “Clear saved data” in the footer deletes exactly that, everything saved locally. There is no project storage on a server; the compressed audio sent to the transcription provider is deleted right after processing.

The on-device mode (opt-in on capable devices) runs entirely on your device, once its model is cached, transcription and caption rendering keep working with the network off. The default free mode sends only compressed audio to the transcription provider, deleted right after processing.

Clear, close-microphone speech is usually very accurate. Proper names, slang, heavy accents and noisy recordings are where mistakes show up. Every word is clickable, so a quick read-through before exporting is enough to fix the stragglers.

Yes. Click any word in the transcript to edit it inline. Words the model was unsure about are flagged with an underline so you know where to look first.

Transcription runs in English today. The engine is Whisper, but the pipeline is currently tuned for English audio, so results in other languages are not reliable yet. Whatever you are captioning, you can correct any word before exporting.

The studio detects where speech actually happens and listens there, and it leans on the voice rather than the music. If the music truly covers the voice, it tells you that instead of inventing words.

Four: an MP4 with the captions burned into the frames, an SRT subtitle file, a WebVTT (.vtt) file, and a plain-text transcript (.txt). MP4 is ready to post; SRT and VTT are for platforms and players that attach subtitles themselves.

In YouTube Studio, open the video, go to “Subtitles”, choose “Add language” and “Upload file”, then pick “With timing” and select your SRT. The same SRT works on Vimeo, LinkedIn and most players that accept subtitle files.

Yes, that's the MP4 export. Pick a caption style first, and the studio renders those captions onto the video frames in your browser while you wait. The result is a normal MP4 you can post anywhere, with no watermark.

Yes. The free studio includes 3 caption styles (Pop Punch, Highlighter and Minimal Clean), exports without a watermark, and no account, 2 minutes of audio per day (resets at 00:00 UTC), videos up to 2 minutes. It is not a trial and it does not expire. All 27 styles preview animated on Free, paid plans unlock applying and exporting them.

Starter ($9 a month) adds 150 caption minutes a month, 4 curated fonts, videos up to 30 minutes, saved projects, cloud storage and 1 brand kit, it keeps the same 3 core caption styles as Free. Pro ($13 a month) raises that to 400 minutes and adds the 6 exclusive caption styles (2 insane builds + 4 clean), unlimited fonts, Boost accuracy (30 server-Whisper minutes a day), Music cleanup and 3 brand kits. Studio ($20 a month) goes to 1,000 minutes, all 27 styles, every font plus your own uploads, unlimited brand kits and a 120-minutes-a-day Boost allowance. Boost minutes draw from the same monthly minutes.

Boost runs Whisper large-v3 on our servers, noticeably better on Hinglish, music-covered vocals and fast or mumbled speech. It is included with the Pro plan (30 minutes a day) and the Studio plan (120 minutes a day); Starter and Free use the standard engine. The counter resets daily at 00:00 UTC and only successfully transcribed audio counts. Flip the “Boost accuracy” toggle in the studio before you drop a video, if Boost is busy or unavailable, the standard engine takes over automatically, so captions are always produced.

Every paid plan is a monthly subscription billed monthly through Lemon Squeezy: it renews automatically, you can cancel anytime, and access continues until the end of the paid period. Minutes reset each billing cycle and unused minutes do not roll over. There is no per-minute pricing.

Usually under a minute. The payment provider confirms the order through a signed webhook, your plan unlocks, and the dashboard shows the purchase under history. If the confirmation is delayed, the webhook retries automatically and if a purchase ever fails to appear, the dashboard's “Restore purchase” button re-syncs it straight from Lemon Squeezy (nothing is charged again).

No. Free exports have never carried a watermark and the paid plans don't add one either.

Check three things: the file is MP4, MOV or WEBM; it's under 500 MB; and the codec is one your browser can play (H.264 MP4 is the safest). Browsers refuse some camera codecs, re-exporting the video as MP4 usually fixes it.

Speed depends on the free server first, opt-in on-device speed depends on your device. A video several minutes long can take a while, keep the tab in the foreground and don't switch away, and consider the smaller model setting for drafts.

Timing is detected from the speech itself, so drift usually comes from very noisy audio or long silences. Re-running the transcription with cleaner audio is the real fix. The text of any word can always be corrected by clicking it.