How DomeScribe handles your files

Convert, compress, extract audio

In your browser (WebAssembly). Nothing uploaded, nothing stored, nothing to delete

Clip Studio editing & export

In your browser. Only the audio your takes cover is sent — and only when you ask for captions

Transcription: uploaded original

Uploaded, transcribed, then deleted after transcription

Transcription: transcript

Encrypted at rest with a per-user key; auto-deleted after 30 days; you can delete sooner

Transcription: audio copy

Kept only for in-app playback for the transcript’s lifetime (30 days max); never downloadable

Pasted platform links

Used only to produce the transcript; the media file is never offered for download

Two kinds of feature, two kinds of handling

The honest answer to “does DomeScribe upload my files?” is: it depends on what you are doing, and the app is built so you can tell. Media tools — convert a video or audio file, compress it, pull the audio track out of a video, and Clip Studio’s framing, captions placement, and vertical render — are browser-local. They use ffmpeg compiled to WebAssembly, a build of the same open-source media engine that runs inside your browser tab, so the file stays on your device from start to finish and the output downloads straight from the browser. There is no server copy to retain or delete because none was ever made.

Transcription is the exception, because speech recognition needs a model larger than a browser tab can run well. The recording is uploaded to DomeScribe’s servers and transcribed by a speech-recognition model; the result is the transcript you see in Transcribe Studio, and everything generated from it — summaries, repurposed drafts, chat answers, clip picks — is produced from that transcript text, not from the media.

Transcription, step by step

When you upload a file to transcribe, it travels to DomeScribe’s servers over an encrypted connection. Video containers are converted to audio server-side, long recordings are split into chunks, and the audio is sent to the speech-recognition model. Once the transcript exists, the uploaded original is deleted. The transcript itself is encrypted at rest with a key that belongs to your account, so transcripts from different users are never encrypted under the same key, and it is auto-deleted 30 days after it was created. You can delete it yourself at any point before then.

One more copy exists, and it is worth being precise about: a compressed audio copy is kept so the player in the app can play the recording beside the transcript — clicking a timestamp seeks to that moment. That copy lives exactly as long as the transcript (30 days at most, or until you delete it) and is never offered for download. When the transcript goes, the audio goes with it.

• Upload → transcribed → original deleted
• Transcript → encrypted per user → auto-deleted after 30 days (or sooner by you)
• Playback audio → kept only for the transcript’s lifetime → never downloadable
• AI outputs (summaries, drafts, chat, clip picks) → generated from the transcript text, stored encrypted with it

Pasted links: a link goes in, a media file never comes out

You can paste a YouTube, TikTok, Facebook, X (Twitter), Vimeo, Twitch, or Dailymotion link in Transcribe Studio. DomeScribe uses the link to produce a transcript — for YouTube videos with captions it can use the existing captions, which means no audio download at all — and the same retention rules apply to the transcript. What DomeScribe will not do is give you the media: there is no download button for platform video or audio, and Clip Studio accepts uploads only, not links. The rule is simple to remember and is enforced everywhere: a link goes in, text comes out, a media file never does.

“I want a transcription tool that doesn’t upload my files”

If the job is converting, compressing, or extracting audio, DomeScribe already is that tool — those run in your browser and nothing is uploaded, which you can verify by loading the tool, switching your device to airplane mode, and running it. If the job is transcription, no honest answer is “no upload”: DomeScribe transcribes on its servers, transiently. What it offers instead is a short, stated chain of custody — upload, transcribe, delete the original, encrypt the transcript under your key, delete it within 30 days, never expose the media — and the option to delete the transcript the moment you have copied what you need.

Journalists, researchers, and anyone handling a sensitive interview can shorten the window further: transcribe, export the transcript (TXT free; DOCX, PDF, JSON on paid plans), then delete the recording in the app. Consider also using the in-browser audio extractor first so that only the audio track, not the video, is ever uploaded.

What DomeScribe does not claim

Plain language cuts both ways, so here is what this page is not saying. DomeScribe does not hold SOC 2, ISO 27001, or HIPAA certifications, and it is not marketed as compliant with any of them. Transcription is not end-to-end encrypted — the server necessarily processes your audio in order to transcribe it; what is encrypted is the connection in transit and the transcript at rest. DomeScribe has no uptime guarantee to quote. And DomeScribe never emails you about the contents of a recording or a transcript. If any of those are requirements for your work, the in-browser tools are the part of DomeScribe you can rely on without a server in the loop.

Questions, answered

Related

• Privacy policy
• In-browser converter — no upload
• Transcribe audio and video to text
• Video converter that works offline

Try it on your own recording

Free to start — 30 minutes of transcription and 5 AI credits, no card required.

Read the privacy policy · See pricing