How to transcribe audio and video by link

Mediata lets you transcribe any publicly available audio or video file by pasting its URL. No downloads, no file conversions — just paste and go.

When to use link-based transcription

Link transcription works best when your recording is already hosted somewhere:

• A Yandex Disk, Google Drive, or Dropbox shared link
• A direct URL to an MP3, MP4, WAV, or other media file
• A podcast episode URL or any publicly accessible media

This is faster than downloading a file to your computer and re-uploading it. Mediata fetches the file directly from the source.

Step-by-step: how it works

1. Open the dashboard and click "New transcription"

In the sidebar, click the New transcription button. You'll see two options: upload a file or paste a link.

2. Choose "From link" and paste the URL

Select the From link option and paste the public URL of your audio or video file. Mediata supports most common formats: MP3, MP4, WAV, FLAC, OGG, WEBM, and more.

3. Confirm language (optional)

If you know the language of the recording, select it. Otherwise, leave it on auto-detect — Mediata identifies the language automatically.

4. Wait for processing

Mediata downloads the file, processes it through the speech recognition engine, and identifies individual speakers. A 30-minute recording typically takes 2-4 minutes to process.

5. Review your transcript

Once ready, you'll see the full transcript with:

• Speaker labels — each person in the recording is identified and color-coded
• Timestamps — every segment has a precise time marker
• Full text — searchable, scrollable, ready for export

6. Analyze in AI chat

Open the AI chat panel next to your transcript. Ask questions like:

• "Summarize the key points from this recording"
• "What decisions were made?"
• "List all action items with responsible people"
• "Find every mention of pricing"

The AI answers based exclusively on the transcript content, with references to specific segments.

Practical tips

Use stable, direct links. Avoid links that require authentication or cookies. If the file is behind a login, download it first and use the file upload option instead.

One recording per request. For the cleanest speaker separation, submit one recording at a time rather than concatenated files.

Check a short sample first. If you plan to process many recordings in bulk, start with one short file to verify quality and speaker detection accuracy.

Supported sources. Any public URL that returns an audio or video file will work. Cloud storage sharing links (Google Drive, Dropbox, Yandex Disk) are fully supported as long as they're set to "anyone with the link."

What you can do after transcription

Once your transcript is ready, Mediata gives you several tools:

• Search — find any word or phrase across the transcript instantly
• AI chat — ask questions, generate summaries, extract action items
• Export — download the transcript as a text file
• Speaker management — rename speakers for clearer records
• Folders — organize transcripts by project, client, or topic

Common AI chat prompts

Here are prompts that work well with transcripts from linked recordings:

• "Summarize key decisions from this meeting"
• "List action items with owners and deadlines"
• "Show all moments where [topic] was discussed"
• "What did [speaker name] say about [subject]?"
• "Create a brief for this interview"
• "Extract all numbers and dates mentioned"