Interview Transcripts

Access, review, and export your interview transcripts—including real-time streaming during active interviews.

Live Transcript

Watch the transcript populate in real-time while an interview is in progress:

1. Navigate to an interview with status IN_PROGRESS
2. The Live Transcript panel shows messages as they happen
3. Both agent and expert messages appear immediately
4. Session indicators show which session is currently active

Multi-Session Support

If an interview has multiple sessions (e.g., a dropped call that was reconnected):

• Session separators clearly mark where each session begins
• Switch between sessions to view different parts of the conversation
• The active session is highlighted

Session Status Indicators

Status	Meaning
Active	Session currently in progress
Completed	Session ended normally
Abandoned	Session ended unexpectedly (call dropped, connection lost)

Accessing Completed Transcripts

Full transcripts are available after an interview completes:

1. Navigate to Interviews
2. Select a completed interview
3. The transcript appears in the left panel

Transcript Features

Speaker Identification

Each entry is labeled with the speaker:

• InsightAgent - The AI interviewer
• Expert - The interview subject

Timestamps

Every entry includes relative timestamps showing time elapsed from the start of the interview (MM:SS format). Click on a timestamp to seek to that point in the audio.

Audio Playback

• Use the audio player to listen to the recording
• Click timestamps to jump to specific moments
• Audio is synchronized with the transcript

Export Options

Available Formats

Format	Best For
Plain Text (.txt)	Simple sharing, notes
JSON	Programmatic access, data processing
PDF	Formal reports, printing

How to Export

1. Open the completed interview
2. Click the Export dropdown
3. Select your desired format
4. The file downloads automatically

API Access

Retrieve transcripts programmatically:

curl -X GET "https://api.insightagent.io/api/interviews/{id}/transcription" \
  -H "Authorization: Bearer <token>"

The response is a JSON array of transcript entries:

[
  {
    "role": "agent",
    "message": "Hello, thank you for joining...",
    "timestamp_secs": 0
  },
  {
    "role": "user",
    "message": "Thanks for having me...",
    "timestamp_secs": 5
  }
]

See API Reference for details.

Improving Accuracy

Use Custom Vocabulary to improve transcription of domain-specific terms, expert names, and acronyms. Vocabulary terms are applied both during real-time recognition and in post-call transcript cleanup.

Privacy & Security

• Transcripts are stored securely in cloud storage
• Access is limited to account members
• Experts cannot access transcripts after the interview
• You control sharing and export