Video and Audio Translation – DekkoDEMS knowledge base

DekkoDEMS can transcribe the speech in an audio or video file and translate it into English at the same time, producing a side-by-side transcript.

Like all transcription in DekkoDEMS, this runs on Dekko EdgeAI, an AI model that executes entirely in your browser, on your own device. The audio, the transcript, and the translation are never sent to a transcription or translation service. Only the resulting transcript is attached to the file inside your case, end-to-end encrypted like everything else in DekkoDEMS. This makes it suitable for sensitive evidence such as interviews, interrogations, bodycam and dashcam footage, and recorded calls.

❗

The transcript and translation are AI-produced drafts. Dekko Edge AI may not be 100% accurate, and machine translation adds a second layer where errors can occur. The original-language column is the verbatim record; the English column is a rough translation. Review both thoroughly against the original recording before relying on them.

If you just want a transcript in the spoken language, without translation, see Video and Audio Transcription.

Supported languages

Translation is always into English, from a spoken language you select: Spanish, French, German, Italian, Portuguese, Dutch, Polish, Russian, Ukrainian, Romanian, Hungarian, Greek, Swedish, Czech, Bulgarian, Slovak, Croatian, Danish, Finnish, Lithuanian, Slovenian, Latvian, Estonian, and Maltese.

Before you start

Permissions - Only the file owner or a case manager with permission to use EdgeAI features can transcribe a file. The action appears on any file DekkoDEMS recognises as audio or video.

One-time model download - The first transcription in a browser downloads the Dekko Edge AI model from Dekko infrastructure, with a progress bar ("Downloading Dekko Edge AI model"). The model is then stored in your browser and reused; later runs start much faster.

💡

The model is cached per browser, so the first run on a new browser, a new device, or after clearing browser data downloads it again.

Transcribing and translating a file

Upload your audio or video file to a case as you normally would. When the upload completes, DekkoDEMS offers to transcribe straight away in the Transcribe media file(s)? dialog. (Already uploaded? Open the file's ··· menu, or More dropdown, and choose Transcribe.)

In the Translate to English dropdown, pick the spoken language of the recording - for example Spanish → English. Leaving it on Don't translate produces a transcript in the original language instead (auto-detected).

Choose Transcribe Now.

Keep the tab open while it runs. The progress card shows the selected language (for example "Language: Es" for Spanish). You can keep working in other tabs, but closing or reloading this tab stops the run and the in-progress transcript is lost.

The bilingual transcript viewer

When the run finishes, the file shows a Transcript badge. The viewer is labelled translated to English and shows two columns per line:

Original - [language] (verbatim). The authoritative transcript of what was actually said.

English ( machine translation - verify) - A working translation to help you navigate and understand the recording.

Everything else works as in a standard transcript:

Click any line to play the recording from that point.

Editable speakers - Rename Speaker 1, Speaker 2, and so on, or merge speakers.

Editable lines - Correct a wrong word or reassign a line; you can also add lines or speakers.

Search the transcript text to jump to a moment.

Detected sounds and a scene/ambience read-out (for example, "Background: quiet — likely indoor"), each an estimate to verify against the recording.

Export the transcript as a plain .txt file, or Save To Case.

💡

Editing the transcript only changes the transcript. It never alters the underlying recording.

Review before you rely on it

A translated transcript has two AI layers to verify: the transcription itself, and the translation on top of it. Before sharing a transcript or using it in any official capacity:

Check the original column against the recording. Misheard or dropped words are possible, especially with crosstalk, noise, accents, or poor audio.

Treat the English column as a rough aid, not a record. Machine translation can miss nuance, idiom, and legal meaning. For evidentiary or court use, follow your organisation's requirements for verified or certified translation.

Confirm speaker labels, sound detections, and the scene read-out. All are model estimates.

Selecting the wrong spoken language produces poor results. If the output looks garbled, re-transcribe with the correct language (this overwrites the previous transcript).

The reviewer confirms; the model only suggests.

Privacy and the audit trail

On-device. Recognition and translation run in your browser, on your hardware. No audio, transcript, or translation text is ever transmitted to a cloud service.

End-to-end encrypted. Files stay encrypted; chunks are decrypted on-device only to process or play.

Non-destructive. Editing a transcript never changes the original recording.

One transcript per file. Re-transcribing (including switching translation on or off) overwrites the previous transcript and resets manual speaker labels.

Auditable, metadata only. Creating, viewing, exporting, and editing a transcript are written to the file's audit log.

Good to know

Formats. Any audio or video file (for example .mp3, .mp4, .webm).

Length. Short recordings transcribe in a single pass; longer files stream a window at a time, so there is no hard length limit, but long recordings take several minutes.

Speed varies by device. A device with a GPU is markedly faster; the fallback runs the same model at the same accuracy, just more slowly.

English only. Translation is currently only into English, from the supported spoken languages above.

FAQ

Is anything uploaded to translate the audio? No. Transcription and translation both run entirely in your browser. The only download is the one-time Dekko Edge AI model, which stays cached for reuse.

Can I translate into a language other than English? Not currently. English is the only translation target.

What if I don't know the spoken language? Run a plain transcription first ("Don't translate" — the language is auto-detected), then re-transcribe with the correct language selected. Note that re-transcribing overwrites the earlier transcript.

Is the English translation accurate enough for court? It is a rough machine translation and is labelled as such in the viewer. Verify it against the recording and follow your organisation's requirements for certified translation in official proceedings.

Can I edit the transcript or translation? Yes. Lines and speaker names are editable in the viewer, and edits never change the original recording.

Can I get the transcript out of DekkoDEMS? Yes - export it as a .txt file from the viewer, or save it to the case.