> ## Documentation Index > Fetch the complete documentation index at: https://docs.cleve.ai/llms.txt > Use this file to discover all available pages before exploring further. # Search & Discovery > Find anything instantly with keyword and semantic AI-powered search ## Overview Cleve's search system helps you find any content across your entire knowledge base instantly. With both traditional keyword search and AI-powered semantic search, you can discover connections and insights you might have missed. ## Opening Search ### Keyboard Shortcut Press `Cmd/Ctrl + /` from anywhere in Cleve to open the search overlay. ### Search Bar Click the search icon in the top navigation or use the search input in any panel. ## Search Types Cleve offers two powerful search modes that work together. ### Full-Text Keyword Search Traditional search that matches exact words and phrases. **How it works:** * Searches through titles, content, and metadata * Matches exact keywords (case-insensitive) * Highlights matching terms in results * Fast and precise **Best for:** * Finding specific quotes or references * Searching for names, dates, or numbers * Locating documents by title **Example queries:** ``` "product roadmap" customer feedback 2024 meeting notes ``` ### Semantic Search AI-powered search that understands meaning, not just keywords. **How it works:** * Your content is embedded as vectors using AI * Your query is also converted to a vector * Finds content with similar meaning * Discovers conceptually related content **Best for:** * Exploring related concepts * Finding content you can't remember exact wording for * Discovering connections between ideas **Example queries:** ``` "strategies for improving team communication" (finds writings about collaboration, meetings, async work, etc.) "dealing with procrastination" (finds content about productivity, motivation, focus) ``` Semantic search works even if the exact keywords don't appear in your content—it understands concepts and context. ## Search Results ### What's Included Search returns results from: * **📝 Writings**: Full documents with titles and content * **💡 Ideas**: Quick captures and notes * **🗂️ Categories**: Folder names and descriptions Each result shows: * **Title or preview** of the content * **Snippet** with highlighted matches * **Creation date** and last modified * **Category** it belongs to * **Relevance score** (semantic search only) ### Result Actions Click a result to: * **Open the writing** in the editor * **View the idea** in context * **Navigate to the category** Right-click for additional options: * Open in new panel * Copy link to content * Share content ## Advanced Search ### Filters Refine your search with filters: * **Type**: Writings only, Ideas only, or All * **Category**: Search within specific categories * **Date Range**: Created or modified within timeframe * **Author**: In shared categories, filter by who created it ### Search Operators Use operators for precise searches: | Operator | Example | Description | | --------------- | --------------------- | ------------------------ | | `"exact match"` | `"customer feedback"` | Find exact phrase | | `author:name` | `author:john` | Filter by author | | `category:name` | `category:blog` | Search specific category | | `created:date` | `created:2024-01` | Filter by creation date | ### Combining Operators ``` "user research" category:product created:2024-02 ``` Finds writings with the exact phrase "user research" in the "product" category created in February 2024. ## Pagination & Load More Search returns the top 20 most relevant results initially. **To see more:** * Scroll to bottom of results * Click **"Load More"** to fetch the next page * Results load incrementally for performance Large result sets are paginated to keep search fast and responsive. ## Search Performance ### Indexing Your content is automatically indexed: * **Instant indexing**: New writings and ideas are searchable immediately * **Background embedding**: Semantic vectors are generated asynchronously * **Incremental updates**: Only changed content is re-indexed ### Speed Optimization Search is optimized for speed: * **\< 100ms** response time for keyword search * **\< 500ms** for semantic search * **Client-side caching** for repeated queries * **Debounced input** to avoid excessive requests ## Search in AI Chat Use search context in AI conversations. ### Enabling Search Context Toggle **"Web Search"** in AI chat settings to: * Search the internet for real-time information * Pull in external sources and citations * Combine your knowledge with current data Toggle **"Writings" or "Ideas"** to search your content as AI context. ### Search-Based Prompts ``` "What did I write about AI in Q1 2024?" → AI searches your writings, summarizes findings "Find ideas related to marketing campaigns" → AI retrieves and organizes relevant ideas ``` ## Privacy & Data ### What's Indexed * **Content**: Full text of writings and ideas * **Metadata**: Titles, dates, categories, authors * **Embeddings**: AI-generated semantic vectors (hashed) ### What's NOT Indexed * **Private writings** are only searchable by you * **Shared content** respects permission levels * **Deleted content** is removed from search immediately ### Search Privacy * All search queries are **private** to your account * Search history is **not logged** or tracked * Semantic embeddings are **one-way hashes** (can't be reversed to original text) ## Mobile Search Search works seamlessly on mobile devices: * **Swipe down** from top to open search * **Voice search**: Speak your query instead of typing * **Quick filters**: Tap to filter by type or category * **Offline search**: Search cached content even without connection ## Use Cases ### Research & Reference ``` Find all mentions of "competitor analysis" across writings → Gather research for a new report ``` ### Content Repurposing ``` Search for "social media tips" → Find past writings to repurpose into new content ``` ### Knowledge Discovery ``` Semantic search: "strategies for growing an audience" → Discover writings about marketing, community building, content creation ``` ### Collaboration ``` Search in shared category: "meeting notes" → Find all team meeting summaries ``` ## Best Practices ### Writing SEO-Friendly Content To make your content more searchable: 1. **Use clear titles**: First H1 becomes the searchable title 2. **Add keywords naturally**: Include terms you'd search for 3. **Structure content**: Headings and lists improve relevance 4. **Link related writings**: Builds semantic connections ### Effective Search Queries **Do:** * Use specific keywords for targeted results * Try semantic search for exploratory queries * Combine filters to narrow results * Use quotes for exact phrases **Don't:** * Search for very generic single words (use filters instead) * Expect semantic search to find exact quotes * Forget to check filters if results seem off ### Organizing for Discoverability * **Consistent category naming**: Makes filtering easier * **Descriptive titles**: Improve search relevance * **Tag with keywords**: Add relevant terms to metadata ## Troubleshooting ### No Results Found **Solutions:** * Check spelling and try alternative phrasing * Remove filters to broaden search * Try semantic search instead of keyword search * Verify content exists and isn't deleted ### Irrelevant Results **Solutions:** * Use exact match with quotes `"like this"` * Add more specific keywords * Use category or date filters * Switch from semantic to keyword search ### Slow Search Performance **Solutions:** * Check internet connection * Clear browser cache * Reduce result limit with filters * Contact support if persistent ## Related Documentation Use search results as context in AI conversations. Learn how to write searchable, well-structured content. Quick captures are also included in search results. Search respects category permissions and sharing settings.