Speech & Voice Transcription

Overview

The Speech & Voice Transcription feature enables you to upload audio recordings (meetings, consultations, interviews) and receive accurate transcriptions with speaker diarization. The transcribed text can then be analyzed using the same AI modes available in Chat.

Key capabilities

• Multi-speaker Recognition: Automatically identifies and labels different speakers in your recording
• High Accuracy Transcription: Powered by Google Cloud Speech-to-Text with legal terminology support
• AI Analysis: Choose between Lawyer mode (focused legal analysis) or Deep Research mode (comprehensive exploration)
• Secure Storage: Audio files are securely stored in your organization's private bucket
• Multiple Languages: Support for Dutch and English audio transcription

Getting Started

1. Access Speech

Navigate to Speech from the top navigation menu.

2. Create a New Speech Conversation

1. Click "New Conversation"
2. Enter a descriptive title (e.g., "Client Meeting - Jan 2025")
3. Select privacy settings:
   • Private: Only you can access
   • Organization: All organization members can view
4. Click "Create"

3. Upload or Record Audio

You have two options:

Upload Existing Audio File

1. Click the Upload tab
2. Drag and drop your audio file or click to browse
3. Configure settings:
   • Language: Select the primary language spoken (Dutch, English, etc.)
   • Number of Speakers: Specify how many people are speaking (improves accuracy)
4. Click "Upload & Transcribe"

Record Live Audio

1. Click the Record tab
2. Grant microphone permissions when prompted
3. Click the Record button to start
4. Pause/resume as needed during your recording
5. Click Stop when finished
6. Configure language and speaker settings
7. Click "Upload & Transcribe"

4. Transcription Process

• Transcription typically takes a few minutes depending on audio length
• You'll see a progress indicator while processing
• The transcription appears automatically when complete
• Each speaker is labeled (Speaker 1, Speaker 2, etc.)

5. Analyze with AI

Once transcribed, you can interact with the content:

1. Select your AI Mode:

   • Lawyer Mode: For specific legal questions and precise citations
   • Deep Research Mode: For broader analysis and context exploration

2. Ask questions about the transcribed content:

   • "Summarize the key legal points discussed"
   • "What are the compliance obligations mentioned?"
   • "Identify any potential legal risks in this conversation"

3. The AI will reference both the transcription and relevant legal sources

Supported Audio Formats

• WebM (recommended for browser recordings)
• MP3
• WAV
• M4A
• FLAC
• Maximum file size: 100MB
• Maximum duration: 3 hours

Best Practices

For Optimal Transcription Quality

1. Clear Audio: Use good quality recording equipment
2. Minimize Background Noise: Record in quiet environments
3. Accurate Speaker Count: Provide the correct number of speakers
4. Proper Language Selection: Choose the primary language spoken

For Better AI Analysis

1. Add Context: Include relevant case numbers or legal topics in the conversation title
2. Specific Questions: Ask targeted questions about the transcription
3. Cross-Reference: Validate AI summaries against the original transcription
4. Use Tags: Organize conversations with tags for easy retrieval

Privacy & Security

• Encrypted Storage: Audio files are encrypted at rest in Google Cloud Storage
• Access Control: Only organization members with appropriate permissions can access
• GDPR Compliant: Audio data is processed in EU regions
• Audit Logs: All access and processing activities are logged

Sync to Chat

You can convert a speech conversation to a regular chat conversation:

1. Open the speech conversation
2. Click the "Sync to Chat" button
3. The transcription will be added as the first message in a new chat conversation
4. Continue with additional questions and AI analysis

Troubleshooting

Transcription Failed

• Check file format: Ensure your audio is in a supported format
• File size: Verify the file is under 100MB
• Audio quality: Poor quality audio may fail to transcribe
• Try again: Some transcription failures are temporary

Poor Transcription Quality

• Speaker count: Adjust the number of speakers
• Language: Ensure correct language is selected
• Audio clarity: Re-record with better equipment if possible
• Background noise: Minimize ambient sounds

Cannot Upload Files

• Permissions: Ensure your organization has file upload enabled
• Storage quota: Check if your organization has available storage
• File format: Verify the audio format is supported
• Browser: Try a different browser if issues persist

Usage is tracked per organization. Check your current usage in Settings → Subscription.