Jinova Compute · Transcribe

Turn spoken content into text.

Upload an audio or video recording. Jinova Compute transcribes the speech and gives you a transcript to read and copy, with timestamps and files to download.

Speech is written out in the language it was spoken in. Jinova does not translate, summarise or label who is speaking.

  • Audio or video
  • Up to 25 minutes
  • Up to 500 MB
  • TXT, JSON, SRT and VTT
Real output

One run of the live workload

An 8-second voice recording went in. This came back.

Input

jinova_compute_transcribe_demo_anika.mp3
0.0 / 8.2 s
01RecordingMP3 · 44.1 kHz · mono · 8.2 s · 129 KB

Output

Language: English · Duration: 8s · Segments: 2 · Characters: 128

Select a segment to play it from its start time.

Transcript
2 segments · 128 characters
Returned files
TXT · JSON · SRT · VTT

Real output from Jinova Compute, shown as returned.

Highlighting follows the start and end times in the returned segments. The workload times each segment, not each word.

How it works

One recording in, text out.

  1. 01

    Upload your recording

    An audio or video file: MP3, WAV, M4A, FLAC, MP4 or MOV.

  2. 02

    See the price

    Jinova reads how long the recording is and shows the price. Nothing runs until you approve it.

  3. 03

    Jinova transcribes

    Everything spoken is written out, in the language it was spoken in.

  4. 04

    Review and download

    Read and copy the transcript, check the timestamps, then download TXT, JSON, SRT or VTT.

Formats and limits

What goes in, what comes out.

One audio or video file per run. The formats and limits are the ones the console shows on the Transcribe page.

Accepted files

  • MP3, WAV, M4A, FLACAudio
  • MP4, MOVVideo

Per recording

  • 25 minLongest recording
  • 500 MBLargest file
  • 1 filePer run

On screen

  • TranscriptRead and copy
  • TimestampsStart of each segment
  • DetailsLanguage, duration, segment and character counts

Downloads

  • TXTThe transcript
  • JSONSegments with start and end times
  • SRTSubtitles
  • VTTSubtitles (WebVTT)

Speech is written out in the language it was spoken in. Jinova does not translate, summarise or label who is speaking.

Uses

Where it helps.

Any time you have a recording and need the words as text.

Meetings

Keep a written record of a recorded meeting that you can search and copy from.

Interviews

Work from the text of an interview instead of scrubbing through the audio.

Podcasts

Publish a written version of an episode alongside the audio.

Recorded research

Turn research sessions into text you can search and review.

Training content

Give recorded lessons and walkthroughs a text version.

Video subtitles

Upload the video, then download the SRT or VTT file to add subtitles to it.

Recordings of other people

If a recording includes other people, you need the legal right to have it transcribed, including any consent the law requires.

Acceptable Use Policy
Jinova Compute

Describe the workload. Jinova handles the compute.

Transcribe is one of the workloads on Jinova Compute. Say what you need in the console or pick the workload, approve the price, and Jinova runs it.

Nothing runs until you approve the price, and you are never charged more than the amount you approve.

Before you upload

The practical questions.

Your files stay yours

You keep ownership of the recordings you upload and of the transcripts Jinova produces for you. We process your files to deliver your result and for nothing else.

An account is required

Uploading and processing happen inside a signed-in account, so your recordings and transcripts are not sitting behind a public link.

Help if something goes wrong

If a job or a charge does not look right, contact us and we will look into it.

How we handle your data

What we collect, how long we keep it and who processes it is set out in our Privacy Policy.

Full detail is in our Privacy Policy and in the Jinova Compute terms shown in the console.

FAQ

Questions about transcribing with Jinova.

What you can upload, what you get back and what it costs.

What file types can I upload?

Audio files in MP3, WAV, M4A or FLAC format, and video files in MP4 or MOV format. Each run transcribes one file.

How large and how long can a recording be?

Up to 500 MB, and up to 25 minutes long.

Can I transcribe a video?

Yes. Upload an MP4 or MOV file and Jinova Compute transcribes the speech in it. You get the transcript and the subtitle files back, not a new video.

What do I get back?

A transcript you can read and copy in the console, with the language it detected, the duration, and a start time for each segment. You can download it as plain text (.txt), as segments with start and end times (.json), and as subtitles (.srt and .vtt).

Does Jinova translate, summarise or identify speakers?

No. Everything spoken is written out in the language it was spoken in. Jinova does not translate, summarise, or label who is speaking.

Can I use the transcript as subtitles?

Yes. The .srt and .vtt downloads are subtitle files with a start and end time for each segment, in the language that was spoken.

Will I know the cost before it runs?

Yes. The console reads how long your recording is and shows the amount to approve before the job runs. Nothing runs until you approve it, and you are never charged more than the amount you approve. Current prices and the terms for Jinova Compute purchases are shown in the console.

Is the demo on this page real?

Yes. It is one run of the live Transcribe workload. The recording you can play is the exact MP3 that went in, and the transcript, timestamps and the four files shown are exactly what came back. Nothing was corrected or edited.

What happens to uploaded files?

You keep ownership of your recordings and transcripts. Jinova processes your files to produce the result you asked for, and you can delete a finished run and its files from the console. What we collect, why, and how long we keep it is set out in our Privacy Policy at /privacy-policy/.

Bring the recording. Take the text.

Jinova Compute does the transcribing. Nothing runs until you approve the price.

Transcribe is one of the workloads on Jinova Compute. It runs in the Jinova console at app.jinovaai.com.

Start transcribingTalk to Jinova