Frequently Asked Questions

Quick answers to the most common DeepTalk questions. For more detailed troubleshooting, see the Common Issues Guide.

Getting Started

What are the system requirements?

Minimum Requirements:

Operating System: Windows 10, macOS 10.14, or Linux (Ubuntu 18.04+)
RAM: 4GB (8GB recommended for large files)
Storage: 2GB free space (more for large content libraries)
CPU: Multi-core processor recommended

For Optimal Performance:

RAM: 8GB+ for large files and projects
Storage: SSD recommended, 10GB+ for content libraries
CPU: 4+ cores for faster processing
Network: Stable internet for external services

Do I need internet connection to use DeepTalk?

Basic Functionality: DeepTalk can work offline for:

Viewing and editing existing transcripts
Basic file management and organization
Exporting content to various formats

Internet Required For:

External transcription services (Speaches, cloud STT)
AI analysis and chat (Ollama, cloud AI services)
Software updates and feature downloads
Community support and documentation access

What audio formats are supported?

Supported Audio Formats:

MP3: Most common, universally supported
WAV: High quality, uncompressed
M4A: Apple format, good compression
FLAC: Lossless compression
OGG: Open source format
AAC: Advanced Audio Coding

Supported Video Formats:

MP4: Most common video format
MOV: Apple video format
AVI: Windows video format
MKV: Matroska video format

File Size Limits:

Maximum file size: 2GB per file
Recommended size: Under 500MB for optimal performance
Duration: No specific limit, but longer files take more time to process

Installation and Setup

How do I install external services?

For Speaches (Enhanced Transcription):

# Option 1: Direct Installation
pip install speaches
speaches serve --port 8000

# Option 2: Docker
docker run -p 8000:8000 speaches/server

For Ollama (AI Analysis):

# macOS/Linux
curl -fsSL https://ollama.ai/install.sh | sh

# Download models
ollama pull llama2
ollama serve

Configuration in DeepTalk:

Open Settings → Services
Set Speaches URL: http://localhost:8000
Set Ollama URL: http://localhost:11434
Test connections and verify functionality

Why can’t DeepTalk connect to external services?

Common Connection Issues:

Service Not Running:

Verify Speaches/Ollama services are started
Check service status in terminal/command prompt
Restart services if necessary

Wrong URL Configuration:

Default Speaches: http://localhost:8000
Default Ollama: http://localhost:11434
Verify URLs in DeepTalk Settings

Firewall/Network Issues:

Check firewall settings allow local connections
Verify no other software using the same ports
Test connectivity with browser or curl

Service Configuration:

Ensure services are configured to accept connections
Check service logs for error messages
Verify correct models are downloaded (for Ollama)

How do I update DeepTalk?

Automatic Updates:

DeepTalk checks for updates automatically
Notification appears when updates are available
Follow prompts to download and install

Manual Update Process:

Check current version: Help → About DeepTalk
Download latest: Visit GitHub releases page
Backup data: Export important projects and settings
Install update: Run installer or replace application
Verify functionality: Test key features after update

Update Best Practices:

Backup before updating: Export critical content
Read release notes: Understand new features and changes
Test after update: Verify everything works as expected
Report issues: Submit bugs if problems occur

Using DeepTalk

How can I improve transcription quality?

Audio Quality Optimization:

Use high-quality recordings: Clear audio produces better transcripts
Minimize background noise: Record in quiet environments
Ensure good microphone placement: Close to speakers, avoid echo
Use proper recording levels: Avoid distortion or very quiet audio

Service Configuration:

Use enhanced services: Speaches provides better accuracy than basic STT
Select appropriate models: Choose models optimized for your content type
Configure language settings: Ensure correct language is selected
Test different services: Compare results across different providers

Post-Processing Improvement:

Review and edit transcripts: Manual correction improves accuracy
Fix speaker labels: Correct speaker identification and names
Add context: Include background information for better understanding
Use consistent terminology: Maintain consistent vocabulary across transcripts

How can I improve analysis quality?

Content Preparation:

High-quality transcripts: Accurate transcripts produce better analysis
Complete speaker information: Proper speaker identification helps analysis
Contextual information: Add relevant background and context
Consistent formatting: Well-formatted transcripts analyze better

AI Service Optimization:

Use appropriate models: Select models suitable for your content type
Configure analysis parameters: Adjust depth and focus areas
Provide clear prompts: Specific analysis requests produce better results
Use context effectively: Include relevant background information

Analysis Customization:

Custom prompts: Create specialized analysis templates
Focus areas: Specify what aspects to emphasize
Industry-specific analysis: Tailor analysis for your domain
Iterative improvement: Refine analysis based on results

How should I organize my projects?

Project Structure Strategy:

Purpose-based: Organize by project goal or outcome
Time-based: Group by date ranges or periods
Team-based: Organize by departments or teams
Content-based: Group by meeting type or subject matter

Effective Organization:

Recommended Project Structure:
├── Active Projects
│   ├── Q1 2024 Planning
│   ├── Product Launch Alpha
│   └── Team Retrospectives
├── Completed Projects
│   ├── Q4 2023 Review
│   ├── Hiring Interviews
│   └── Training Sessions
└── Templates and Standards
    ├── Meeting Templates
    ├── Analysis Templates
    └── Export Formats

Best Practices:

Consistent naming: Use standardized naming conventions
Descriptive titles: Include date, purpose, and participants
Regular maintenance: Archive completed projects, update active ones
Tag systematically: Use consistent tagging for easy discovery

What’s the difference between search modes?

Vector Search Mode:

How it works: Finds similar content based on semantic meaning
Best for: Discovering related concepts and themes
Speed: Fast response times
Results: Ranked by similarity with confidence scores

Keyword Search Mode:

How it works: Finds exact word or phrase matches
Best for: Finding specific quotes or precise information
Speed: Very fast
Results: Exact matches with context

AI Chat Mode:

How it works: Combines search with AI interpretation
Best for: Natural language questions and analysis
Speed: Slower but more comprehensive
Results: Conversational responses with explanations

When to Use Each:

Keyword search: Finding specific quotes, names, or facts
Vector search: Exploring themes, discovering related content
AI chat: Complex questions requiring interpretation and analysis

Advanced Features

How do I set up automated workflows?

Basic Automation Setup:

Define triggers: What events should start automation
Configure actions: What should happen when triggered
Set parameters: Customize behavior and outputs
Test workflow: Verify automation works as expected

Common Automation Examples:

Daily reports: Automatic export of daily transcript summaries
Project updates: Regular analysis and reporting for ongoing projects
Quality assurance: Automatic review and validation of processed content
Integration sync: Automatic export to external systems

Advanced Automation:

Custom scripts: Develop specialized automation scripts
API integration: Connect with external workflow systems
Conditional logic: Complex decision-making in automated processes
Error handling: Robust error recovery and notification

How do I integrate with other tools?

Common Integrations:

Project Management:

Export action items: Send tasks to Jira, Asana, Trello
Sync timelines: Update project schedules based on meeting content
Track decisions: Monitor decision implementation and outcomes

Document Management:

Archive transcripts: Store in SharePoint, Google Drive, Box
Generate reports: Create documents in standard company formats
Version control: Maintain document history and changes

Communication Platforms:

Share insights: Post summaries to Slack, Teams channels
Notify stakeholders: Automated updates via email or messaging
Collaboration: Enable team discussion around transcript content

Integration Methods:

API connections: Programmatic integration with business systems
Export automation: Scheduled or triggered content distribution
Webhook integration: Real-time notifications and updates
Manual workflows: Streamlined manual processes for common tasks

Can I customize analysis prompts?

Yes, DeepTalk supports extensive analysis customization:

Custom Prompt Creation:

Industry-specific analysis: Tailor prompts for your business domain
Role-based perspectives: Generate insights for different stakeholder groups
Outcome-focused analysis: Emphasize specific types of results
Quality standards: Apply consistent analysis criteria

Prompt Categories:

Executive summaries: High-level insights for leadership
Technical analysis: Detailed examination of technical discussions
Action item extraction: Focus on tasks and follow-up items
Risk assessment: Identify potential issues and concerns

Template Management:

Save reusable prompts: Create libraries of analysis templates
Share with team: Distribute standard analysis approaches
Version control: Maintain prompt history and improvements
Performance tracking: Monitor effectiveness of different prompts

Troubleshooting

DeepTalk won’t start or crashes frequently

Common Startup Issues:

Check system requirements: Verify minimum hardware/software requirements
Free up resources: Close other applications, restart computer
Check permissions: Ensure DeepTalk has necessary file system access
Review logs: Look for error messages during startup

Crash Prevention:

Update software: Ensure latest version of DeepTalk
Monitor resources: Watch memory and CPU usage during operation
Manage file sizes: Break up very large files into smaller segments
Regular maintenance: Clear temporary files and restart periodically

Processing takes too long or fails

Performance Optimization:

For Transcription:

Audio quality: Better quality audio processes faster and more accurately
File size: Smaller files process faster; consider splitting large files
Service selection: Try different transcription services for speed/quality balance
System resources: Ensure adequate RAM and CPU available

For AI Analysis:

Model selection: Faster models for routine analysis, comprehensive models for detailed work
Content scope: Analyze specific sections rather than entire long transcripts
Prompt efficiency: Use focused, specific analysis prompts
Service configuration: Optimize AI service settings for your hardware

Processing Failures:

Check service connectivity: Verify external services are responding
Review file format: Ensure supported audio/video format
Monitor system resources: Avoid processing during high system load
Retry with smaller content: Test with shorter files to isolate issues

My exports don’t look right

Common Export Issues:

Formatting Problems:

Template selection: Choose appropriate template for your needs
Content selection: Verify all desired elements are included
Format compatibility: Ensure target format supports required features
Preview before export: Check preview to identify issues early

Missing Content:

Selection settings: Verify content selection includes all desired elements
Permission issues: Ensure export has access to all referenced content
Template configuration: Check that template includes all required sections
Processing status: Ensure all content has completed processing

Quality Improvement:

Template customization: Modify templates to better fit your needs
Content preparation: Improve source content quality before export
Format optimization: Choose formats that best support your content type
Review and iterate: Test different options to find optimal settings

I can’t find content I know I uploaded

Content Discovery:

Search Strategies:

Use multiple search terms: Try different keywords and phrases
Check all projects: Content might be in unexpected projects
Review recent items: Check Recent section on Home page
Search by metadata: Use date, speaker, or tag filters

Organization Check:

Project assignments: Verify content is assigned to expected projects
Archive status: Check if content has been archived
Permission settings: Ensure you have access to view content
Folder structure: Review project and folder organization

Recovery Options:

Check trash/deleted items: Content might be in trash folder
Review import history: Check upload logs for processing status
Backup recovery: Restore from backup if content was accidentally deleted
Contact support: Get help if content appears to be lost

Getting Help

Where can I get additional support?

Community Resources:

GitHub Discussions: https://github.com/michael-borck/deep-talk/discussions
GitHub Issues: https://github.com/michael-borck/deep-talk/issues (for bug reports)
Documentation: Complete guides in the docs/ directory

Self-Help Resources:

Common Issues Guide: Detailed troubleshooting for frequent problems
User Guide: Comprehensive feature explanations and usage guidance
Tutorials: Step-by-step learning workflows
Features Guide: In-depth capability documentation

Before Requesting Help:

Search existing issues: Check if your problem has been reported
Try troubleshooting steps: Follow relevant troubleshooting guides
Gather information: Include error messages, system info, steps to reproduce
Test reproducibility: Verify the issue occurs consistently

How do I report bugs or request features?

Bug Reports:

Check existing issues: Search GitHub issues for similar problems
Gather details: Error messages, system information, steps to reproduce
Create clear report: Use GitHub issue template with all relevant information
Provide examples: Include sample files or screenshots if helpful

Feature Requests:

Describe use case: Explain the problem you’re trying to solve
Propose solution: Suggest how the feature might work
Consider alternatives: Think about different approaches to the same goal
Engage community: Discuss with others who might benefit

Effective Communication:

Be specific: Provide detailed, actionable information
Be respectful: Remember that development is volunteer work
Be patient: Complex issues take time to investigate and resolve
Follow up: Provide additional information if requested

Still need help? Check the Common Issues Guide for detailed troubleshooting, or visit our GitHub Discussions for community support.