Skip to main content
Revolutionize content discovery with context-aware search capabilities that go beyond simple keyword matching. The Intelligent Search feature leverages advanced semantic understanding to help users find relevant content and communities based on meaning and context, dramatically improving user engagement and content discoverability.

​
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 - 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?”
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 - 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.
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
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
Major Languages: Arabic, Swahili, Hausa, Yoruba, Amharic, Somali, Malagasy, Xhosa, Zulu, AfrikaansRegional Languages: Kinyarwanda, various other regional dialects
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

Characteristics:
  • Perfect semantic alignment with search intent
  • High lexical similarity
  • Ideal for strict relevance requirements
Example: Searching “red fruits” returns “strawberry recipes” and “cherry cultivation”
Characteristics:
  • Strong semantic relevance
  • Good lexical matches
  • Recommended for most use cases
Example: Searching “red fruits” returns “apple varieties” and “raspberry farming”
Characteristics:
  • Moderate relevance
  • Some contextual connection
  • Useful for broader discovery
Example: Searching “red fruits” returns “healthy snacks” and “fruit nutrition”
Characteristics:
  • Weak relevance
  • Tangential connections
  • May include unrelated content
Example: Searching “red fruits” might return “colorful vegetables” or “fruit trees”

​
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

  • Low Threshold (0.3)
  • High Threshold (0.8)
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)
Use Case: Content discovery, exploring related topics

​
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

  • 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
  • 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
  • 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