Settings Reference

Complete guide to all Astronox settings and configuration options.

Accessing Settings

• Button: Click ⚙️ icon in top-right corner
• Keyboard: Cmd/Ctrl + ,
• From chat: Say "Open settings"

General Settings

Access Mode

Controls how Astronox handles potentially risky operations.

Safe Mode (Recommended)

What it does:

• Asks for confirmation before risky operations
• Shows preview of what will happen
• Provides approve/deny buttons
• Explains consequences

When confirmations appear:

• File deletions
• Bulk file operations (>10 files)
• System setting changes
• Process termination
• Shell commands (except safe reads)
• Irreversible actions

Best for:

• New users learning Astronox
• Critical work environments
• When working with important files
• Understanding what AI is doing

Example:

⚠️ About to delete 15 files. Confirm?
Files: temp1.txt, temp2.txt, ...
[✓ Proceed] [✗ Cancel]

Full Mode (Advanced)

What changes:

• Most operations execute immediately
• Only very high-risk actions ask confirmation
• Faster workflow
• More trust in AI's judgment

Still asks confirmation for:

• Bulk deletions (>50 files)
• System-critical operations
• Disabling safety features
• Permanent data loss

Best for:

• Experienced users
• Trusted workflows
• Automation scripts
• When speed matters

⚠️ Warning: Be cautious. Review AI responses carefully.

Theme

Visual appearance of Astronox.

Available themes:

• Dark (default): Dark background with teal accents, optimized for low-light environments and eye comfort
• Light: Light background with dark text, optimized for bright environments and daytime use

How to Change:

1. Open Settings (⚙️ button or Cmd/Ctrl + ,)
2. Find the Theme section
3. Select Light or Dark using radio buttons
4. Your preference is saved automatically

What's Affected:

• Chat interface and message bubbles
• Sidebar and conversation list
• All buttons and interactive elements
• Input composer area
• Health checks panel
• Modals and overlays
• All text colors and backgrounds

Light Mode Features:

• Carefully optimized contrast for readability
• Softer teal accents that work with light backgrounds
• Lighter borders and backgrounds
• Perfect for daytime productivity

Dark Mode Features:

• Reduces eye strain in low-light environments
• Vibrant teal accents on dark background
• Better battery life on OLED displays
• Default theme for most users

Future Customization (planned):

• Accent color selection
• Font size adjustment
• Compact/comfortable density
• Auto theme switching based on system settings

Language & Region

Language:

• Currently: English only
• Future: Multi-language support planned

Region:

• Date format preferences
• Time zone settings
• Number formatting

Startup Preferences

On Launch:

• Open last conversation (default: checked)
• Check for updates automatically
• Restore window size and position
• Launch with Windows/macOS (default: disabled)
• Minimize to tray when closed (default: enabled)

Default Access Mode:

• Select Safe or Full as startup default

Personalization

User Name

What it does:

• AI uses your name in responses
• More natural, personalized conversation
• Optional field

Example:

• Without: "I can help with that."
• With: "I can help with that, Alex."

Personal Information

What it does:

• Additional context about you for AI
• Job role, interests, preferences
• Helps AI provide more relevant responses

Examples:

• "I'm a Python developer working on web apps"
• "I prefer functional programming patterns"
• "I work on macOS and deploy to Linux servers"

Privacy:

• Stored locally only
• Never shared with third parties
• Can be cleared anytime

API Keys & Providers

Provider Grid

Select your preferred LLM provider from a 3×3 grid in Settings:

Row 1: Auto · Gemini · OpenAI
Row 2: Claude · Azure OpenAI · MintAI (Pro)
Row 3: NVIDIA (dev mode only)

Each provider has its own API key field and model selection.

Google Gemini

Setup:

1. Get free API key from Google AI Studio
2. Select Gemini from provider grid
3. Paste key in API Key field
4. Select model (Flash, Flash Lite, or Pro)
5. Click "Save"

Free Tier:

• Flash/Flash Lite: 60 req/min, 1500 req/day
• Pro: 10 req/min, 500 req/day

Models:

• Gemini 2.5 Flash (default, recommended)
• Gemini 2.5 Flash Lite (fastest)
• Gemini 2.5 Pro (best quality)

OpenAI (GPT-5.2)

Setup:

1. Get API key from OpenAI Platform
2. Select OpenAI from provider grid
3. Paste key in API Key field
4. Select model (GPT-5.2 or GPT-5.2-Codex)
5. Click "Save"

Models:

• GPT-5.2 (default, all-purpose)
• GPT-5.2-Codex (code optimized)

Pricing: ~$0.01–0.05 per request (see API Keys for details)

Claude (Anthropic)

Setup:

1. Get API key from Anthropic Console
2. Select Claude from provider grid
3. Paste key in API Key field
4. Select model (Haiku 4.5 or Sonnet 4.5)
5. Click "Save"

Models:

• Claude Haiku 4.5 (default, fast & affordable)
• Claude Sonnet 4.5 (more powerful reasoning)

Pricing: ~$0.005–0.02 per request

Azure OpenAI

Setup:

1. Deploy Azure OpenAI resource from Azure Portal
2. Select Azure OpenAI from provider grid
3. Paste your Endpoint URL
4. Paste your API Key
5. Select model (GPT-5.2 or GPT-5.2-Codex)
6. Click "Save"

Models:

• GPT-5.2 (default)
• GPT-5.2-Codex

Best for: Enterprise, compliance, regional deployment

MintAI (Pro)

Setup:

• Automatic for Pro subscribers
• No manual configuration needed
• Key fetched from managed feed

Requirements:

• Active Pro subscription
• Internet connection

Auto Mode

What it does:

Select Auto from the provider grid to let Astronox choose the best provider based on task complexity:

1. Simple queries → Gemini Flash (fast, free)
2. Moderate tasks → Gemini Pro (balanced)
3. Complex/reasoning → Claude Sonnet 4.5 (best reasoning)
4. Code generation → GPT-5.2-Codex (code optimized)
5. Fallback → Automatic retry if provider fails

Benefits:

• ✨ Best model for each task
• 💰 Optimized cost
• ⚡ Optimized speed
• 🎯 Optimized quality

API Key Management

Security:

• Keys encrypted locally
• Never transmitted except to chosen provider
• Not stored in cloud
• Can be cleared anytime

Actions:

• Show/Hide: Toggle key visibility
• Test Connection: Verify key works
• Clear Key: Remove stored key
• Regenerate: Get new key from provider

Troubleshooting:

• Invalid key: Double-check from provider dashboard
• Connection failed: Check internet, firewall
• Rate limited: Wait or upgrade to paid tier

Model Selection

Choose default model for new conversations.

Available models:

Gemini:

• Gemini 2.5 Flash (default, best balance)
• Gemini 2.5 Flash Lite (fastest)
• Gemini 2.5 Pro (best quality)

OpenAI:

• GPT-5.2 (default, all-purpose)
• GPT-5.2-Codex (code optimized)

Claude:

• Claude Haiku 4.5 (fast, affordable)
• Claude Sonnet 4.5 (advanced reasoning)

Azure:

• GPT-5.2 (default)
• GPT-5.2-Codex

MintAI (Pro):

• Devstral 2 (Mistral AI)

Per-conversation:

• Can switch models mid-conversation
• New messages use newly selected model
• Previous messages retain original model's responses

See Model Selection Guide for detailed comparison.

Temperature

Controls AI creativity and randomness in responses.

Range: 0.0 - 2.0
Default: 0.15

Lower values (0.0 - 0.5):

• More deterministic and focused
• Consistent, predictable responses
• Better for code generation and facts

Medium values (0.5 - 1.0):

• Balanced creativity and accuracy
• Good for general conversation

Higher values (1.0 - 2.0):

• More creative and varied
• Less predictable
• Better for brainstorming

Recommended:

• Code/Technical: 0.1 - 0.3
• General use: 0.15 (default)
• Creative writing: 0.7 - 1.2

Advanced Settings

Memory System

Controls how Astronox remembers information across conversations.

Enable Memory

• Store facts and preferences (default: enabled)

What gets stored:

• User preferences (languages, tools, workflows)
• Project context (names, locations, tech stacks)
• Custom facts you ask to remember
• Learned patterns from interactions

What does NOT get stored:

• API keys or passwords
• Full conversation transcripts
• Sensitive personal data (unless you ask)

Memory Categories

Built-in categories:

• Personal (preferences, work style)
• Projects (active work, codebases)
• Technical (languages, frameworks, tools)
• Custom (user-defined)

Management:

• View all stored facts: "Show me what you remember"
• Clear category: "Forget everything about work projects"
• Clear all: Settings → Advanced → Clear Memory

Clipboard History

Enable Clipboard Tracking:

• Track clipboard history (default: enabled)

Settings:

• History size: 50 items (configurable: 10-200)
• Retention: Until app restart (not persisted)
• Privacy: Local only, never uploaded

What's tracked:

• Text copied to clipboard
• Timestamps
• Source application (future)

What's NOT tracked:

• Passwords (if detected)
• Credit card numbers (pattern-matched)
• Content from password managers

Clear clipboard history: Settings → Advanced → Clear Clipboard

Process Guard

Enable Process Guard:

• Prevent accidental termination of critical processes

Protected processes:

• System processes (always)
• Long-running builds/compilations (optional)
• Database servers (optional)
• Your customized list

Behavior:

• Extra confirmation required
• Shows process details before termination
• Cannot be bypassed in Safe Mode

Automation Settings

Save Location:

• Default: ~/Library/Application Support/Astronox/automations/
• Customizable in future

Auto-save prompts:

• Suggest saving useful scripts automatically

Execution safety:

• Review scripts before first run
• Sandbox execution (future)

Privacy & Data

Conversation Storage

Location:

• macOS: ~/Library/Application Support/Astronox/chat_sessions/
• Windows: %APPDATA%\Astronox\chat_sessions\

Format: JSON files (one per conversation)

Includes:

• Messages (yours and AI's)
• Attachments (base64-encoded)
• Timestamps
• Model used

Actions:

• Export conversation: Settings → Advanced → Export (future)
• Clear all conversations: Permanent deletion
• Backup: Manually copy chat_sessions folder

What Gets Sent to AI Providers

Sent:

• Your messages
• Attached files/images
• Relevant system context (OS, paths when needed for tasks)
• Conversation history (for context)

Never sent:

• Unrelated files
• API keys
• Passwords
• Full file system scan
• Clipboard history (unless specifically requested)

Analytics & Telemetry

Currently collected:

• None (zero telemetry in current version)

Future (opt-in):

• Crash reports (anonymous)
• Feature usage (anonymous)
• Performance metrics

Your control:

• All telemetry will be opt-in
• Can be disabled anytime
• No tracking without permission

Performance

Streaming Responses

• Stream AI responses in real-time (default: enabled)

Benefits:

• Faster perceived response time
• See progress as AI works
• Cancel mid-stream if needed

Disable if:

• Prefer complete responses
• Network connection unreliable

Stream Thinking

• Show AI's thinking process in real-time (default: enabled)

What it does:

• Displays reasoning steps as AI plans
• Shows tool selection and decision-making
• Provides transparency into AI behavior

Benefits:

• Understand how AI approaches tasks
• Catch potential issues early
• Learn AI's problem-solving process

Disable if:

• Prefer cleaner, simpler output
• Only care about final results

Health Checks

• Monitor system health (default: enabled)

What it monitors:

• API connection status
• Local system resources
• App performance metrics

Benefits:

• Early warning for issues
• Better error messages
• Improved reliability

Disable if:

• Minimal resource usage preferred
• Running on low-spec hardware

Message History Limit

Per conversation:

• Default: 100 messages
• Configurable: 20-1000

Why limit:

• Manages token usage
• Keeps conversations focused
• Improves performance

When limit reached:

• Oldest messages truncated (not deleted)
• Can export before truncation

Update Settings

Automatic Updates:

• Check for updates on launch
• Notify when new version available
• Auto-download updates (future)

Update channel:

• Stable (default): Tested releases only
• Beta (future): Early access to features

Manual check: Settings → About → Check for Updates

Debug & Logging

Enable Debug Mode:

• Show detailed logs (default: disabled)

When enabled:

• Verbose console output
• API request/response logging
• Performance metrics
• Error stack traces

Log location:

• macOS: ~/Library/Logs/Astronox/
• Windows: %APPDATA%\Astronox\logs\

⚠️ Warning: Debug logs may contain sensitive data. Disable for normal use.

Settings by Use Case

For Beginners

Recommended:

• Access Mode: Safe
• Model: Gemini Flash
• Memory: Enabled
• Process Guard: Enabled
• Streaming: Enabled

For Power Users

Recommended:

• Access Mode: Full
• Model: Gemini Pro or Devstral 2
• Memory: Enabled
• Clipboard: Enabled (200 items)
• Streaming: Enabled

For Developers

Recommended:

• Access Mode: Full
• Model: Devstral 2 or Gemini Pro
• Memory: Enabled
• Debug: Enabled (when troubleshooting)
• Auto-save scripts: Enabled

For Privacy-Conscious Users

Recommended:

• Access Mode: Safe
• Memory: Disabled (or curate carefully)
• Clipboard: Disabled
• Telemetry: Disabled (when available)
• Review what's sent to AI before sending

Importing/Exporting Settings

Export Settings (Future)

Location: Settings → Advanced → Export Configuration

Exports:

• All settings (except API keys)
• Memory facts
• Saved automations
• Custom categories

Format: JSON file

Use cases:

• Backup configuration
• Transfer to new computer
• Share setup with team

Import Settings (Future)

Location: Settings → Advanced → Import Configuration

Merges:

• Settings overwrite existing
• Memory facts added (not replaced)
• Automations merged

Resetting Settings

Reset to Defaults

Location: Settings → Advanced → Reset to Defaults

Resets:

• All settings to factory defaults
• Does NOT delete conversations
• Does NOT delete memory facts
• Does NOT remove API keys

Use when:

• Settings corrupted
• Want fresh start
• Troubleshooting issues

Clear All Data

Location: Settings → Advanced → Clear All Data

⚠️ Danger: Permanent deletion!

Deletes:

• All conversations
• All memory facts
• All saved automations
• Clipboard history
• API keys

Preserves:

• App settings
• Update preferences

Use when:

• Selling/transferring computer
• Starting completely fresh
• Privacy requirement

Settings File Location

For manual editing (advanced users only):

macOS:

~/Library/Application Support/Astronox/settings.json

Windows:

%APPDATA%\Astronox\settings.json

⚠️ Warning: Manual editing may break settings. Backup first!

Keyboard Shortcuts in Settings

Troubleshooting Settings Issues

Settings Won't Save

Try:

1. Check file permissions on settings.json
2. Close and reopen Astronox
3. Reset to defaults, then reconfigure
4. Delete settings.json (will reset all)

Settings Disappeared

Solution:

• Settings file corrupted
• Check backup: settings.json.backup
• Or reset to defaults

Can't Change Access Mode

Possible causes:

• Process guard preventing change
• Debug mode interference
• Restart app

Next: Learn about Access Modes & Safety in detail.