Key Features
Semantic Understanding
Advanced AI-powered search that understands context and meaning, not just keywords
Multi-Language Support
Support for 100+ languages with consistent search quality across different locales
Intelligent Scoring
Hybrid scoring system combining lexical (30%) and semantic (70%) relevance for optimal results
This feature requires activation. Contact our support team to enable Intelligent Search for your network.
Search Parameters
Core Parameters
Query Parameter
Query Parameter
query - The search term that can be:- Single words: “coffee”
- Phrases: “best coffee shops”
- Sentences: “Where can I find good coffee in downtown?”
- Questions: “What are the benefits of organic coffee?”
Minimum Score Filtering
Minimum Score Filtering
minScore - Filter results by relevance threshold (0.0 to 1.0):- 0.0 - 0.3: Broad results, may include loosely related content
- 0.4 - 0.6: Balanced relevance, good for general use
- 0.7 - 1.0: Highly relevant results only, strict filtering
Language Support
Language Support
language - Specify search language for optimal results:- Auto-detection available for most queries
- Manual specification improves accuracy
- Support for 100+ languages (see full list below)
Supported Languages
Our intelligent search supports over 100 languages, ensuring global accessibility and consistent search quality across different locales.European Languages
European Languages
Western Europe: English, French, German, Spanish, Italian, Portuguese, Dutch, Swedish, Norwegian, Danish, Finnish, IcelandicEastern Europe: Russian, Polish, Czech, Slovak, Hungarian, Romanian, Bulgarian, Croatian, Serbian, Slovenian, Estonian, Latvian, LithuanianOthers: Greek, Albanian, Basque, Catalan, Galician, Irish, Welsh, Maltese, Luxembourgish
Asian Languages
Asian Languages
East Asia: Chinese, Japanese, Korean, MongolianSoutheast Asia: Thai, Vietnamese, Indonesian, Malay, Tagalog, Burmese, Khmer, LaoSouth Asia: Hindi, Bengali, Tamil, Telugu, Marathi, Gujarati, Kannada, Malayalam, Punjabi, Urdu, Nepali, Sinhala, Assamese, OriyaCentral/West Asia: Arabic, Persian, Turkish, Hebrew, Kurdish, Azerbaijani, Kazakh, Kirghiz, Uzbek, Turkmen, Tajik
African Languages
African Languages
Major Languages: Arabic, Swahili, Hausa, Yoruba, Amharic, Somali, Malagasy, Xhosa, Zulu, AfrikaansRegional Languages: Kinyarwanda, various other regional dialects
Other Languages
Other Languages
Constructed Languages: EsperantoHistorical Languages: Latin, SanskritRegional Variants: Norwegian Nynorsk, Western Frisian, Occitan, Breton, Corsican, Faroese
Scoring Algorithm
Our intelligent search uses a sophisticated hybrid scoring system that combines multiple factors to deliver the most relevant results.Score Composition
Lexical Score (30%)
Traditional keyword matching that identifies exact or similar text matches in content, titles, and descriptions.
Semantic Score (70%)
AI-powered understanding of meaning and context, enabling discovery of relevant content even without exact keyword matches.
Score Range & Interpretation
Score Range: 0.9 - 1.0 (Excellent Match)
Score Range: 0.9 - 1.0 (Excellent Match)
Characteristics:
- Perfect semantic alignment with search intent
- High lexical similarity
- Ideal for strict relevance requirements
Score Range: 0.7 - 0.8 (Very Good Match)
Score Range: 0.7 - 0.8 (Very Good Match)
Characteristics:
- Strong semantic relevance
- Good lexical matches
- Recommended for most use cases
Score Range: 0.5 - 0.6 (Good Match)
Score Range: 0.5 - 0.6 (Good Match)
Characteristics:
- Moderate relevance
- Some contextual connection
- Useful for broader discovery
Score Range: 0.0 - 0.4 (Low Match)
Score Range: 0.0 - 0.4 (Low Match)
Characteristics:
- Weak relevance
- Tangential connections
- May include unrelated content
Network-Level Configuration
Advanced filtering and scoring parameters can be configured at the network level through the Network Settings API. This requires admin permissions.
Practical Examples
Example 1: Broad vs. Focused Search
Query: “Red Fruits”Results Include:
- 🍎 Apples (direct match)
- 🍓 Strawberries (direct match)
- 🍌 Bananas (loosely related - fruit category)
- 🥕 Carrots (color match)
- 🍕 Tomato recipes (red color, fruit technically)
Current Limitations
Known Constraints:
- User Activity: Recent user behavior and preferences are not factored into result ranking
- Real-time Suggestions: Type-ahead search suggestions are not available across all API endpoints
- Content Freshness: Results prioritize relevance over recency (can be configured)
- Multilingual Queries: Mixed-language queries may have reduced accuracy
Implementation Guides
Ready to implement intelligent search in your application?Intelligent Search Posts
Learn how to implement semantic search for posts with filtering and customization options
Intelligent Search Communities
Discover how to enable intelligent community search with category and tag filtering
Best Practices
Query Optimization
Query Optimization
- Use natural language queries for best semantic understanding
- Avoid overly technical jargon unless searching technical content
- Consider user intent when setting minimum score thresholds
- Test different query formulations for optimal results
Performance Considerations
Performance Considerations
- Implement appropriate pagination for large result sets
- Cache frequently searched terms when possible
- Use debouncing for real-time search implementations
- Monitor search performance and adjust thresholds accordingly
User Experience
User Experience
- Provide clear feedback about search scope and filters
- Explain scoring thresholds to users when appropriate
- Implement fallback to keyword search for no results
- Consider showing related suggestions for low-result queries