Transcript Upload
Upload or paste interview transcripts to get the same AI-powered insights as video interviews — no recording required.
Transcript Upload
Don't have a video or audio file? No problem. Deck now lets you upload a raw interview transcript and get the same AI-powered synthesis, themes, and insights as a full video interview — in seconds.
Supported Formats
Deck automatically detects and parses transcript formats. You can also manually select a format from the dropdown if detection is incorrect:
| Format | How to export | Auto-detected? |
|---|---|---|
WebVTT (.vtt) | Export from Zoom, Teams, Google Meet, or most transcription tools | Yes |
Google Meet (.txt) | Download the transcript from a Google Meet recording | Yes |
| Plain Text | Any unformatted text separated by blank lines — no speaker labels or timestamps required | Yes (3+ paragraphs) |
You can also paste the transcript text directly — Deck handles format detection automatically.
Tip: Plain text transcripts are now automatically detected when your text contains three or more paragraphs separated by blank lines. If detection picks the wrong format, you can still manually select it from the dropdown.
How to Upload a Transcript
- Navigate to Manage → User Interviews
- Click Add Interview — a two-step upload wizard opens
- In the Upload step, select the Transcript tab
- Choose how you want to add your transcript:
- Upload a file — drag and drop a
.vtt,.txt, or.docxfile - Paste text — paste the transcript directly into the text area
- Upload a file — drag and drop a
- Deck automatically detects the format and shows a live preview of the parsed segments
- Enter a required Interview Name and Interview Date, and optionally a customer email
- Click Next to proceed to the Review step
- Edit the transcript segments as needed in the transcript editor (rename speakers, fix text, add or remove segments)
- Click Upload Interview to start AI processing
Live Preview
As you paste or upload, Deck immediately parses the transcript and shows you:
- Speaker segments — who said what, in order
- Format detected — VTT, Google Meet, or Plain Text
- Segment count — how many speaker turns were found
If the format isn't detected correctly, you can manually select it from the dropdown.
Editing the Transcript
Once the preview shows your parsed segments, you can edit them directly before submitting:
- Change speaker — use the dropdown on any segment to reassign it to a different speaker
- Edit text — click the text area on any segment to correct transcription errors
- Delete a segment — remove any segment that doesn't belong in the interview
This is especially useful with Plain Text transcripts, where all content is initially assigned to a single speaker. You can add new speaker labels and re-attribute segments to the correct participants before processing.
Note: Edits apply only to what Deck processes — your original file is not modified.
Interview Details
During the Upload step, fill in your interview metadata:
| Field | Required | Description |
|---|---|---|
| Interview Name | Yes | A label for this interview (e.g. "User Research Session #3") |
| Interview Date | Yes | When the interview took place |
| User Email | No | Your customer's email address |
The Interview Name and Interview Date must be provided before you can advance to the Review step. The customer email is optional and can be added later from the interview detail page.
What Happens Next
Once you submit, Deck's AI agents process the transcript the same way they process video interviews:
- Insight extraction — key themes and pain points are identified
- Quote mapping — verbatim quotes are linked to each insight
- Theme tagging — insights are grouped into your existing themes or new ones are created
- Summary generation — a narrative summary of the interview is produced
Processing typically completes in under 2 minutes.
Tips for Best Results
- Minimum length: Transcripts must be at least 100 words and 10 lines. Very short excerpts won't be processed.
- Maximum length: Up to 50,000 words are supported.
- Speaker labels: Make sure speaker names are consistent throughout the transcript (e.g., always "Jane Smith" not sometimes "Jane" and sometimes "Interviewer").
- Clean transcripts: Remove any introductory notes or boilerplate that isn't part of the conversation itself.
Transcript vs. Video — What's Different?
| Video/Audio | Transcript | |
|---|---|---|
| Transcription | Automatic via AssemblyAI | You provide the transcript |
| Video clips | Available (jump to moment) | Not available |
| Processing time | ~5 minutes | ~2 minutes |
| Insights & themes | ✓ | ✓ |
| Speaker identification | ✓ | ✓ |
| Quote extraction | ✓ | ✓ |