Settings & Preferences
Customize your ai.KMITL experience with personal preferences, model defaults, and feature configurations. This guide covers all available settings.
Accessing Settings
Click your avatar → Settings
Settings Overview
ai.KMITL settings are organized into several categories:
Profile
- Profile Information: View your name, email, and account details
- Active Sessions: Manage devices logged into your account
- Session Security: Revoke sessions and sign out of other devices
Appearance
- Display Mode: Choose between Light and Dark theme
- Chat Width: Select Normal or Wider chat message display
Customization
- Your Name: Set how the AI should address you personally
- AI Personality: Define the AI's communication style and behavior
- Additional Context: Provide background information about yourself for more relevant responses
AI Options
- AI Memory: Configure Supermemory integration for cross-conversation context
- MCP Servers: Set up Model Context Protocol servers for extended capabilities
- Web Search Provider: Choose between TinyFish (default), Brave, Tavily, Firecrawl, or Serper for web searches
- Search Sources: Toggle whether to include source links by default in search results
Models
- Available Models: View all text and image generation models you can access
- Custom Models: Add your own custom model configurations
- Model Status: See which models are available through system, BYOK, or custom providers
Providers
- Core Providers: Configure API keys for OpenAI, Anthropic, Google, Groq, and OpenRouter
- Custom Providers: Add your own OpenAI-compatible API endpoints
- Search Providers: Set up API keys for TinyFish, Brave, Tavily, Firecrawl, and Serper
Attachments
- File Management: View all files you've uploaded
- Storage Usage: See total storage used and file count
- File Actions: Download or delete uploaded files
Profile Settings
Editing Your Name
- Go to Settings → Profile
- Click Edit next to your name
- Enter your preferred name
- Click Save
Your name appears in your profile and can be used by the AI for personalization.
Managing Active Sessions
View all devices logged into your account:
Each session shows:
- Device type (Desktop, Mobile, Tablet)
- Browser information
- Last seen timestamp
- IP address
- Current session indicator
Actions:
- Sign Out: End your current session
- Revoke: End another device's session
- Revoke All Other Sessions: Sign out of all devices except current
Security Best Practice
Regularly review your active sessions and revoke any you don't recognize.
Appearance Settings
Display Mode
Choose your preferred theme:
Light Mode:
- Bright, clean interface
- Better for well-lit environments
- Easier on eyes in daylight
Dark Mode:
- Reduced eye strain in low light
- Battery saving on OLED screens
- Modern, sleek appearance
The theme applies immediately across the entire interface.
Chat Width
Control how wide chat messages display:
Normal (Default):
- Standard width for optimal readability
- Centered layout with comfortable margins
- Best for most use cases
Wider:
- Extended width for more content
- Better for code and tables
- Maximizes screen real estate
Customization Settings
Your Name
Tell the AI what to call you in conversations.
Example:
Your Name: "Alex"
AI: "Hi Alex! How can I help you today?"Max length: 100 characters
AI Personality
Shape how the AI communicates with you.
Examples:
Professional:
"Be formal, concise, and business-focused. Use professional
language and avoid casual expressions."Friendly & Casual:
"Be warm, conversational, and approachable. Use casual
language and feel free to add personality to responses."Teaching Style:
"Explain concepts like a patient teacher. Break down complex
topics, use analogies, and check my understanding."Max length: 2,000 characters
Additional Context
Provide background information that helps the AI give more relevant responses.
What to include:
- Your profession or field of study
- Programming languages you use
- Areas of interest
- Learning goals
- Work context
Example:
I'm a computer science student learning web development.
I work with React, TypeScript, and Node.js. I prefer detailed
explanations with code examples. I'm interested in AI and
machine learning.Max length: 2,000 characters
Personalization Tips
- Be specific about your needs and preferences
- Update context as your needs change
- The AI applies these settings to every conversation
- Settings work across all AI models :::
AI Options Settings
AI Memory (Supermemory)
Coming Soon
Supermemory integration is currently under active development and will be available soon!
Enable AI to remember information across conversations.
Setup:
- Get API key from Supermemory
- Enter key in AI Options
- Toggle Enable Supermemory
- Save changes
How it works:
- AI stores important information you share
- Retrieves relevant context in future conversations
- Works across different chat threads
- Respects your privacy settings
MCP Servers
Connect Model Context Protocol servers for extended capabilities.
Configuration:
- Click Add MCP Server
- Enter server name and endpoint
- Optionally add API key
- Save configuration
Use cases:
- Connect to internal databases
- Integrate custom tools
- Access private APIs
- Extend AI capabilities
Web Search Provider
Choose which service powers web searches:
| Provider | Status | Description |
|---|---|---|
| TinyFish | Server/BYOK | Free, agent-oriented search (default) |
| Brave | BYOK | Privacy-focused, fast results |
| Tavily | BYOK | AI-optimized search |
| Firecrawl | BYOK | Deep web crawling |
| Serper | BYOK | Google Search API wrapper |
Server: Included with system quota
BYOK: Requires your own API key (unlimited)
If your selected provider has no key available (for example Brave without your own key), searches use TinyFish automatically.
To change:
- Go to AI Options → Web Search Provider
- Click your preferred provider
- Selection saves automatically
Search Sources
Control whether search results include source links:
Enabled (Default):
- Shows URLs and citations
- Verify information easily
- Follow up with original sources
Disabled:
- Cleaner responses
- Focus on answers
- Less clutter
Models Settings
Viewing Available Models
See all models you can access:
Model types:
- Text generation models (GPT, Claude, Gemini, etc.)
- Image generation models (DALL-E, Imagen, etc.)
For each model you see:
- Model name and version
- Capabilities (vision, reasoning, function calling, etc.)
- Active provider (System, BYOK, Custom)
- Quota cost multiplier
Provider priority:
- Your BYOK keys (if configured)
- Custom providers (your endpoints)
- OpenRouter (if you have a key)
- System keys (included access)
Understanding Model Cards
Each model displays:
Name & Icon:
- Provider logo
- Model name
- Short name (if applicable)
Status Badge:
- Available (green) - Ready to use
- Unavailable (gray) - Need BYOK or custom provider
Capabilities:
- 👁️ Vision - Can process images
- 🧠 Reasoning - Advanced problem-solving
- 🔧 Function Calling - Can use tools
- 📄 PDF - Can read PDF files
- 🎚️ Effort Control - Adjustable reasoning depth
Provider Info:
- Shows which service provides access
- "System" = Included with quota
- "OpenAI BYOK" = Your API key
- "Custom: MyEndpoint" = Your custom configuration
Adding Custom Models
Add your own OpenAI-compatible models:
Go to Settings → Models
Click Add Custom Model
Fill in details:
- Model ID: Technical identifier
- Display Name: How it appears in UI
- Provider: Select custom provider endpoint
- Context Length: Max tokens (optional)
- Max Output: Max response tokens (optional)
- Capabilities: Select supported features
Click Save
Example:
Model ID: llama-3-70b
Display Name: Llama 3 70B
Provider: My Local LLM
Context Length: 8000Editing & Deleting Models
Edit a model:
- Find the model card
- Click Edit button (pencil icon)
- Modify settings
- Save changes
Delete a model:
- Click Delete button (trash icon)
- Confirm deletion
- Model removed from your list
System Models
You cannot edit or delete built-in system models. Only custom models you've added can be modified.
Providers Settings
Core Providers (BYOK)
Add API keys for major AI providers:
OpenAI
- Models: GPT-5, GPT-4o, o3, DALL-E
- Get Key: platform.openai.com
- Pricing: See BYOK Guide
Anthropic
- Models: Claude Sonnet, Opus, Haiku
- Get Key: console.anthropic.com
- Pricing: Pay per token
Google AI
- Models: Gemini 2.5 Pro, Flash, Imagen
- Get Key: ai.google.dev
- Pricing: Often includes free trial
Groq
- Models: Llama, Mixtral (ultra-fast)
- Get Key: console.groq.com
- Pricing: Competitive rates
OpenRouter
- Models: Access to 100+ models
- Get Key: openrouter.ai
- Pricing: Varies by model
Adding a Provider Key
- Go to Settings → Providers
- Find the provider card
- Click Edit (pencil icon)
- Toggle Enable [Provider] on
- Paste your API key
- Click Save
Your key is:
- Encrypted at rest
- Never logged or shared
- Only used for your requests
- Deletable anytime
Rotating API Keys
Update an existing key:
- Click Edit on provider
- Click Rotate Key
- Enter new API key
- Save changes
Old key is immediately replaced.
Custom Providers
Add your own OpenAI-compatible endpoints:
Use cases:
- Local LLM servers (Ollama, LM Studio)
- Self-hosted models
- Internal company APIs
- Proxy services
Setup:
Click Add Custom Provider
Enter details:
- Name: Provider identifier
- Display Name: UI label
- Endpoint: API base URL
- API Key: Authentication token (if needed)
Toggle Enable
Save configuration
Example:
Name: local-ollama
Display Name: Ollama (Local)
Endpoint: http://localhost:11434/v1
API Key: (leave empty for local)Search Provider BYOK
Add API keys for web search services:
Brave Search
- Fast, privacy-focused search
- brave.com/search/api
TinyFish
- Free search and page extraction built for AI agents
- tinyfish.ai
Tavily
- AI-optimized search results
- tavily.com
Firecrawl
- Deep web content extraction
- firecrawl.dev
Serper
- Google Search API wrapper
- serper.dev
Setup same as core providers:
- Toggle enable
- Paste API key
- Save
Attachments Settings
Viewing Your Files
See all files you've uploaded to ai.KMITL:
File information:
- File name
- File type (image, PDF, document, etc.)
- File size
- Upload date
Storage summary:
- Total files count
- Total storage used
- Storage breakdown by type
Managing Files
Download a file:
- Find the file in list
- Click Download button
- File opens in new tab
Delete a file:
- Click Delete button (trash icon)
- Confirm deletion
- File permanently removed
File Deletion
Deleted files cannot be recovered. Messages referencing deleted files will still show, but attachments will be unavailable.
File Types Supported
ai.KMITL accepts:
Images:
- JPG, JPEG
- PNG
- GIF
- WebP
Documents:
- TXT
- Markdown
Archives:
- ZIP
- And more
Limits:
- Max file size: 50MB
- Max files per message: 10
Settings Tips & Best Practices
Personalization
- Set your name for a personal touch
- Define AI personality to match your communication style
- Add context about your work/studies for better responses
Security
- Review active sessions regularly
- Revoke unknown sessions immediately
- Use unique API keys for ai.KMITL
Cost Management
- Start with system quota (included)
- Add BYOK when you need unlimited access
- Monitor provider usage in their dashboards
Model Selection
- System models are always available
- BYOK unlocks all models
- Custom models for specialized needs
File Organization
- Regularly review and delete unused files
- Keep storage tidy
- Re-upload files when needed
Troubleshooting
Settings not saving?
- Check internet connection
- Try refreshing the page
- Sign out and sign back in
- Clear browser cache
API key not working?
- Verify key is correct (no extra spaces)
- Check provider account has credits
- Ensure billing is set up at provider
- Try rotating the key
Models unavailable?
- Add BYOK key for that provider
- Check provider key is enabled
- Verify account has credits
- Try OpenRouter as backup
Files not uploading?
- Check file size (max 50MB)
- Verify file type is supported
- Try a different browser
- Check internet connection
Privacy & Security
All settings are:
- Encrypted at rest and in transit
- Private to your account only
- Secure following best practices
- Deletable anytime you want
See Privacy Policy for details.
Next Steps
- Configure API Keys (BYOK) for unlimited access
- Check Usage Quotas to monitor consumption
- Review Data Privacy for security details
- Read Personalization Guide for customization tips