> ## Documentation Index > Fetch the complete documentation index at: https://docs.qwox.ai/llms.txt > Use this file to discover all available pages before exploring further. # Search > Find anything instantly with AI-powered semantic search ## Overview Qwox uses advanced **semantic search** powered by AI. Unlike traditional keyword search, Qwox understands the **meaning** of your query and finds relevant information even if the exact words don't match. ## How It Works ### Semantic AI Search Qwox converts your search query into an AI-powered representation that captures its semantic meaning. This allows Qwox to find conceptually similar content, even when the exact words differ. **Example:** | You Search | What Qwox Finds | | ----------------- | -------------------------------------------------------------------- | | "car problems" | Notes about "vehicle issues", "auto repair", "mechanic appointments" | | "happy moments" | Memories tagged with "celebration", "achievement", "good news" | | "workout routine" | Notes containing "exercise plan", "fitness schedule", "gym sessions" | ### Intelligent Ranking Qwox combines two factors to rank search results: 1. **Relevance**: How closely the meaning matches your query 2. **Recency**: How recent the content is You can adjust this balance with sort modes (see below). ## Search Methods ### Via Chat The most natural way to search: ```text theme={null} "Find my marketing notes" "What do I know about Sarah?" "Show me my workout routine" "Search for doctor appointments" ``` Qwox will automatically search your memories and display relevant results. ### Via Library Use the search bar in the Library screen: * **Mobile**: Tap search icon → Type query → View results * **Web**: Search bar at top → Type query → Results appear instantly ### Via AI Agent Ask Qwox to search for specific information: ```text theme={null} "What's John's email address?" "When did I last go to the dentist?" "Find my flight confirmation" ``` The AI will search your memories and provide a natural language answer. ## Search Features ### Smart Filters Narrow your search results with filters: #### Type Filter ```text theme={null} "Find notes about marketing" "Search events next week" "Show me uploaded files" ``` **Supported Types:** * `note` - Text notes * `event` - Calendar events * `file` - Uploaded documents * `all` - Everything (default) #### Collection Filter ```text theme={null} "Search in Project Alpha collection" "Find notes in Work collection" ``` Collections are automatically extracted from your notes based on projects, people, locations, and organizations mentioned. #### Tag Filter ```text theme={null} "Find items tagged important" "Search notes with tag #work" ``` Multiple tags can be combined (matches any tag). ### Sort Modes Control how results are ranked: | Mode | Description | Focus | | ----------------------- | ------------------- | -------------------------- | | **Relevance** (default) | Best semantic match | Prioritizes meaning match | | **Newest** | Most recent first | Prioritizes creation date | | **Recently Updated** | Recently modified | Prioritizes last edit date | **Example:** ```text theme={null} "Search for meeting notes, sort by newest" "Find project documents, sort by relevance" ``` ### Result Quality Qwox automatically filters out low-quality matches to show you only relevant results. * **Precise Mode**: Fewer, highly relevant results * **Broad Mode**: More results, may include loosely related items ## Web Search Integration Qwox can search the internet when you ask: ```text theme={null} "Search the web for Python tutorials" "Find recent news about AI" "Look up weather in Paris" ``` **Features:** * **Professional Search**: Enterprise-grade search technology * **Language Support**: Searches in your preferred language * **Location-Aware**: Results tailored to your location * **Organic Results Only**: No ads, just relevant content Web search is separate from memory search. Qwox automatically determines whether to search your memories or the web based on your query. ## Advanced Features ### Context-Aware Search Qwox considers: * **Recent Conversations**: What you've been discussing * **Current Time**: "yesterday", "last week" are understood * **Related Information**: Connects people, projects, and events ### Multi-Language Content Qwox can find content regardless of language: * Notes in Turkish, English, French, etc. * Searches work across all languages * Results ranked by semantic similarity, not language match While Qwox can search across languages, results are most accurate when your query language matches your content language. ### Fuzzy Matching Qwox handles: * **Typos**: "meetting" finds "meeting" * **Synonyms**: "doctor" finds "physician" * **Related Concepts**: "fitness" finds "gym", "workout", "exercise" ## Search in Action ### Example 1: Finding Contact Info **Query:** "What's Sarah's phone number?" **What Qwox does:** 1. Analyzes the meaning of "Sarah phone number" 2. Searches memories for semantic matches 3. Finds note: "Sarah Johnson - 555-0123" 4. Returns result ranked by relevance ### Example 2: Finding Events **Query:** "When is my next dentist appointment?" **What Qwox does:** 1. Searches events with "dentist appointment" 2. Filters for future dates 3. Sorts by date (ascending) 4. Returns: "Dr. Smith - March 15, 2:00 PM" ### Example 3: Project Research **Query:** "Find all notes about Project Alpha" **What Qwox does:** 1. Semantic search for "Project Alpha" 2. Applies collection filter (if Project Alpha is a collection) 3. Returns all related notes, sorted by relevance ## Search Tips "Find my dentist appointment with Dr. Smith" is better than "find appointment" Ask questions naturally: "What's John's email?" instead of "John email search" If you can't find something, try synonyms: "doctor" → "physician", "meeting" → "appointment" Narrow results with type, collection, or tag filters for faster discovery Use "Newest" to find recent items, "Relevance" for best semantic match ## Performance ### Search Speed * **Typical Response**: Under 1 second * **First Search**: May take 1-3 seconds (system initialization) * **Optimization**: Intelligent caching for faster subsequent searches ### Scalability * **Unlimited Memories**: No limit on searchable content * **Efficient Transfer**: Only relevant results downloaded (not all memories) * **Cloud-Powered**: Advanced AI processing in the cloud ## Limitations The following features are **NOT currently supported**: * ❌ **Boolean Operators** (AND, OR, NOT) - Use natural language instead * ❌ **Exact Phrase Search** (quotes) - Semantic search finds similar meaning * ❌ **Date Range Filters** - Use natural language like "last month" or "this year" * ❌ **Field-Specific Search** (title:, tag:) - Use filters instead * ❌ **Search History** - Previous searches are not saved * ❌ **Cross-Language Search** - Query and content should be in same language for best results ## Troubleshooting * Try broader terms or synonyms * Check if content exists in your memories * Lower the similarity threshold (if using API) * Verify you're searching the correct type (notes vs events) * Be more specific in your query * Use filters (type, collection, tags) * Try "Relevance" sort mode * Increase similarity threshold * First search after inactivity may be slow (cold start) * Subsequent searches are faster (warm function) * Check your internet connection * Try "Newest" or "Recently Updated" sort mode * Check if item was actually saved * Refresh the library view ## Related Guides Understand how information is stored Organize content for better searchability Search via natural conversation