10 - Settings & Configuration Reference

⚙️ Complete Settings Documentation
⏱️ Time Estimate: 15 minutes
📋 What You’ll Learn: All settings explained, configuration best practices, advanced options

Accessing Settings
LLM & Processing Settings
License Management
Automation Settings
Display Preferences
Data Management
Advanced Settings
Settings Import/Export

Accessing Settings

Opening Settings Page

Method 1: Header Icon

Click ⚙️ icon in header (top-right)

Method 2: Keyboard Shortcut (Planned)

Ctrl+, (Windows/Linux)
Cmd+, (macOS)

Settings Page Layout

┌────────────────────────────────┐
│ Settings                       │
├────────────────────────────────┤
│ Sections:                      │
│ • LLM & Processing             │
│ • License Management           │
│ • Automation                   │
│ • Display Preferences          │
│ • Data Management              │
│ • Advanced                     │
└────────────────────────────────┘

💡 Pro Tip: Settings auto-save as you make changes. No “Save” button needed for most options.

LLM & Processing Settings

Feature: F007 - Two-Stage LLM Configuration

Configure separate providers for transcription and analysis.

Transcription Settings (Audio → Text)

┌─────────────────────────────────┐
│ Transcription LLM Settings      │
├─────────────────────────────────┤
│ Provider:   [Ollama         ▼]  │
│ Model:      [whisper:base   ▼]  │
│ Endpoint:   [localhost:11434]   │
│ API Key:    [____________] (N/A)│
│                                 │
│ [Test Connection]               │
└─────────────────────────────────┘

Provider Options:

Ollama (local, free, private)
OpenAI (cloud, fast, paid)
Gemini (cloud, cost-effective)

Model Selection:

For Ollama:

whisper:tiny - Fastest, basic accuracy
whisper:base - Balanced (recommended)
whisper:small - Higher accuracy
whisper:medium - Best accuracy
whisper:large - Maximum accuracy

For OpenAI:

whisper-1 - Only option (state-of-the-art)

For Gemini:

Auto-detected Gemini Audio model

Endpoint:

Default: http://localhost:11434
Custom: For remote Ollama instances

API Key:

Required for: OpenAI, Gemini
Not needed for: Ollama

Analysis Settings (Text → Insights)

┌─────────────────────────────────┐
│ Analysis LLM Settings           │
├─────────────────────────────────┤
│ Provider:   [OpenAI         ▼]  │
│ Model:      [gpt-4o-mini    ▼]  │
│ Endpoint:   [api.openai.com]    │
│ API Key:    [sk-proj-******]    │
│                                 │
│ [Test Connection]               │
└─────────────────────────────────┘

Provider Options:

Ollama (local, free)
OpenAI (cloud, fast, expensive)
Gemini (cloud, cost-effective)

Model Selection:

For Ollama:

llama3.1:latest - Recommended
llama3.1:70b - Best quality (requires 48GB RAM)
Custom models (see below)

For OpenAI:

gpt-4o - Best quality, expensive
gpt-4o-mini - Balanced (recommended)
gpt-3.5-turbo - Fast, cheaper

For Gemini:

gemini-1.5-flash - Fast, cheap (recommended)
gemini-1.5-pro - Best quality

Cost Comparison:

Provider	Model	Cost (10-page transcript)
Ollama	llama3.1	Free
OpenAI	gpt-4o-mini	~$0.02
OpenAI	gpt-4o	~$0.08
Gemini	1.5 flash	~$0.01
Gemini	1.5 pro	~$0.05

Custom Models

Adding custom models:

1. Click "Manage Custom Models"
2. Select provider
3. Enter model name
4. Click "Add"

Example custom models:
- llama3.1:custom-tuned
- gpt-4o:2024-08-06
- gemini-exp-1114

Use cases:

Fine-tuned domain-specific models
Beta/experimental models
Self-hosted model endpoints

Testing Connections

Test Connection button:

Verifies API key is valid
Checks endpoint accessibility
Confirms model availability

Success indicators:

✅ Connection successful
   Model: gpt-4o-mini
   Latency: 234ms

Failure indicators:

❌ Connection failed
   Error: Invalid API key
   [Check Settings]

License Management

Feature: F014 - LemonSqueezy Integration

┌─────────────────────────────────┐
│ License Information             │
├─────────────────────────────────┤
│ Status:   ✅ Active              │
│ Plan:     Lifetime Professional │
│ Email:    user@example.com      │
│ Activated: Oct 16, 2024         │
│                                 │
│ [Deactivate License]            │
│ [Upgrade Plan]                  │
└─────────────────────────────────┘

Activation

First-time activation:

Enter license key
Click “Activate”
Wait for validation
✅ Success confirmation

License key format: XXXX-XXXX-XXXX-XXXX

Deactivation

When to deactivate:

Moving to new device
Selling/giving away device
Troubleshooting license issues

Process:

Click “Deactivate License”
Confirm deactivation
License becomes available for other device
App switches to trial mode

⚠️ Note: You can reactivate on same device anytime.

Upgrades

Upgrade paths:

Trial → Lifetime
Basic → Professional
Professional → Enterprise

Process:

Click “Upgrade Plan”
Redirects to purchase page
Complete payment
Enter new license key
✅ Plan upgraded

Automation Settings

Auto-Transcribe

┌─────────────────────────────────┐
│ Automation Preferences          │
├─────────────────────────────────┤
│ ☑ Auto-transcribe after         │
│   recording                     │
│                                 │
│ ☑ Auto-analyze after            │
│   transcription                 │
│                                 │
│ ☐ Auto-export to cloud          │
│   (coming soon)                 │
└─────────────────────────────────┘

Auto-transcribe after recording:

✅ Enabled: Immediately processes audio
❌ Disabled: Manual “Transcribe Audio” button

Use cases:

Enable for: Quick workflows, trusted providers
Disable for: Cost control, selective processing

Auto-analyze after transcription:

✅ Enabled: Generates insights automatically
❌ Disabled: Manual “Start Analysis” button

Use cases:

Enable for: End-to-end automation
Disable for: Review transcript first, batch processing

💡 Pro Tip: Disable both for maximum control, enable both for fully automated workflow.

Display Preferences

Theme Settings

┌─────────────────────────────────┐
│ Appearance                      │
├─────────────────────────────────┤
│ Theme:  ○ Light                 │
│         ● Dark                  │
│         ○ System                │
│                                 │
│ Accent: [🟡 Yellow (default) ▼] │
└─────────────────────────────────┘

Theme options:

Light: White backgrounds, dark text
Dark: Dark backgrounds, light text (default)
System: Matches OS preference

Accent colors:

🟡 Yellow (default)
🔵 Blue
🟢 Green
🟣 Purple
🔴 Red

┌─────────────────────────────────┐
│ Customize Widget Order          │
├─────────────────────────────────┤
│ Drag to reorder:                │
│                                 │
│ ☰ Meeting Metadata              │
│ ☰ Decision Flowchart            │
│ ☰ Concept Mind Map              │
│ ☰ Action Matrix                 │
│ ☰ Sentiment Arc                 │
│ ☰ Participation Heatmap         │
│                                 │
│ [Reset to Default]              │
└─────────────────────────────────┘

Customization:

Drag handles (☰) to reorder
Changes apply to all transcripts
Per-project ordering (planned)

Recommended orders:

For executives:

1. Action Matrix
2. Decision Flowchart
3. Sentiment Arc
4. Meeting Metadata

For project managers:

1. Action Matrix
2. Participation Heatmap
3. Decision Flowchart
4. Concept Mind Map

Data Management

Export & Backup

┌─────────────────────────────────┐
│ Data Management                 │
├─────────────────────────────────┤
│ Export All Data:                │
│ [Create Full Backup]            │
│                                 │
│ Import Data:                    │
│ [Restore from Backup]           │
│                                 │
│ Storage Usage:                  │
│ Database:    245 MB             │
│ Audio:       6.8 GB             │
│ Models:      1.2 GB             │
│ Total:       8.2 GB             │
│                                 │
│ [Cleanup Old Files]             │
└─────────────────────────────────┘

Export options:

Individual transcripts (JSON, PDF)
Project archives (ZIP)
Full backup (complete database)

Import options:

Single transcripts
Project restores
Complete data restore

👉 Learn More: See 09_DATA_MANAGEMENT_GUIDE.md

Storage Cleanup

Cleanup wizard:

Select cleanup criteria:
- Audio older than X months
- Deleted project files
- Unused Whisper models
Preview space to reclaim
Confirm deletion
✅ Space freed

Advanced Settings

API Endpoints (Custom)

┌─────────────────────────────────┐
│ Advanced Configuration          │
├─────────────────────────────────┤
│ Custom Endpoints:               │
│                                 │
│ Ollama:                         │
│ [http://localhost:11434____]    │
│                                 │
│ OpenAI:                         │
│ [https://api.openai.com____]    │
│                                 │
│ Gemini:                         │
│ [https://generativelanguage.]  │
│                                 │
│ [Reset to Defaults]             │
└─────────────────────────────────┘

Use cases:

Remote Ollama server
OpenAI-compatible proxies
Custom API gateways

Processing Configuration

┌─────────────────────────────────┐
│ Processing Options              │
├─────────────────────────────────┤
│ Timeout: [120 seconds      ▼]   │
│ Retries: [3 attempts       ▼]   │
│                                 │
│ ☐ Enable verbose logging        │
│ ☐ Cache API responses           │
│ ☑ Validate JSON schemas         │
└─────────────────────────────────┘

Timeout:

Default: 120 seconds
Increase for: Large transcripts, slow models
Decrease for: Quick failures

Retries:

Default: 3 attempts
Exponential backoff between retries
Prevents transient errors

Verbose logging:

Enables debug-level logs
Useful for troubleshooting
Logs saved to app directory

Database Optimization

┌─────────────────────────────────┐
│ Database Maintenance            │
├─────────────────────────────────┤
│ Last optimized: 3 days ago      │
│                                 │
│ [Vacuum Database]               │
│ [Rebuild Indexes]               │
│ [Verify Integrity]              │
│                                 │
│ ⚠️ Backup before optimizing     │
└─────────────────────────────────┘

Operations:

Vacuum:

Reclaims unused space
Defragments database
Improves performance

Rebuild Indexes:

Refreshes search indexes
Fixes corruption
Speeds up queries

Verify Integrity:

Checks for corruption
Reports issues
Safe to run anytime

Settings Import/Export

Exporting Settings

Use cases:

Transfer settings to new device
Share team configurations
Backup preferences

Process:

Settings → Advanced → Export Settings
    ↓
Save settings.json file

Contains:

{
  "llm_settings": {
    "transcription_provider": "ollama",
    "analysis_provider": "openai",
    "models": {...}
  },
  "display_preferences": {
    "theme": "dark",
    "widget_order": [...]
  },
  "automation": {
    "auto_transcribe": true,
    "auto_analyze": true
  }
}

⚠️ Note: API keys are NOT included for security.

Importing Settings

Process:

Settings → Advanced → Import Settings
    ↓
Select settings.json file
    ↓
Choose what to import:
☑ LLM providers (no API keys)
☑ Display preferences
☑ Automation settings
☐ Advanced options
    ↓
[Import]

Conflict resolution:

Imported settings overwrite existing
API keys must be re-entered manually
Backup created automatically

Settings Best Practices

For Privacy

Recommended:

Transcription:  Ollama (whisper:base)
Analysis:       Ollama (llama3.1)
Auto-transcribe: Enabled
Auto-analyze:   Enabled

Result: 100% local, zero cloud calls

For Speed

Recommended:

Transcription:  OpenAI (whisper-1)
Analysis:       OpenAI (gpt-4o-mini)
Auto-transcribe: Enabled
Auto-analyze:   Enabled

Result: 10-20 second processing, ~$0.05/transcript

For Cost

Recommended:

Transcription:  Ollama (whisper:base)
Analysis:       Gemini (1.5 flash)
Auto-transcribe: Enabled
Auto-analyze:   Disabled (manual review)

Result: Only pay for analysis, ~$0.01/transcript

For Accuracy

Recommended:

Transcription:  OpenAI (whisper-1)
Analysis:       GPT-4o
Auto-transcribe: Enabled
Auto-analyze:   Enabled

Result: Best quality, ~$0.10-0.15/transcript

Troubleshooting Settings

Settings Not Saving

Symptoms:

Changes revert after reload
“Failed to save” errors

Solutions:

✅ Check disk space (need > 100MB)
✅ Verify database is writable
✅ Close other Selfoss instances
✅ Restart application

API Keys Not Working

Symptoms:

“Invalid API key” errors
Connection tests fail

Solutions:

✅ Remove extra spaces from key
✅ Verify key is active in provider account
✅ Check spending limits/quota
✅ Try regenerating key

Settings UI Issues

Symptoms:

Dropdowns not working
Buttons unresponsive

Solutions:

✅ Clear browser cache (if using web view)
✅ Restart application
✅ Reset settings to defaults
✅ Check for updates

Next Steps

🎉 You’ve mastered Selfoss settings!

Recommended Actions:

⚙️ Configure for your use case - Choose speed/privacy/cost balance
🧪 Test connections - Verify all providers work
💾 Export settings - Backup your configuration
🎨 Customize display - Set widget order and theme
📊 Monitor usage - Track costs if using cloud

Advanced Topics:

Team settings templates
Automated configuration deployment
API proxy setup for enterprise
Custom model fine-tuning

⚙️ Fine-tune Selfoss to work exactly how you want.

10 - Settings & Configuration Reference

10 - Settings & Configuration Reference

Table of Contents

Accessing Settings

Opening Settings Page

Settings Page Layout

LLM & Processing Settings

Feature: F007 - Two-Stage LLM Configuration

Transcription Settings (Audio → Text)

Analysis Settings (Text → Insights)

Custom Models

Testing Connections

License Management

Feature: F014 - LemonSqueezy Integration

Activation

Deactivation

Upgrades

Automation Settings

Auto-Transcribe

Display Preferences

Theme Settings

Widget Ordering (F021)

Data Management

Export & Backup

Storage Cleanup

Advanced Settings

API Endpoints (Custom)

Processing Configuration

Database Optimization

Settings Import/Export

Exporting Settings

Importing Settings

Settings Best Practices

For Privacy

For Speed

For Cost

For Accuracy

Troubleshooting Settings

Settings Not Saving

API Keys Not Working

Settings UI Issues

Next Steps

Recommended Actions:

Advanced Topics: