ai_similar_content
The AI Similar Content module brings real-time, AI-powered suggestions to Drupal content editors, helping them discover similar content while typing. It simplifies content creation by improving discoverability through semantic similarity search.
Features
What is the basic functionality?
The module performs real-time similarity searches across text fields, textareas, and WYSIWYG editors (CKEditor 4 and 5). As the user types, the system suggests similar content items.
What unique features does enabling this project add?
- Real-time AI-powered suggestions based on semantic similarity
- Uses embeddings to understand contextual relevance
- Seamless integration with the Milvus vector database for efficient nearest-neighbor search
- Configurable settings including debounce delay, similarity threshold, and content types
- Clean, responsive UI to display suggestions beneath input fields
When and why would someone use this module?
This module is perfect for content teams aiming to improve content reuse, reduce redundancy, and ensure consistent messaging across large repositories. It allows authors to instantly find relevant existing content without having to leave the editing interface.
What use cases are there?
- Suggesting related blog posts during article writing
- Recommending similar products or services in commerce setups
- Enhancing editorial workflows with content discovery tools
Post-Installation
Once installed, the module can be configured by visiting:
/admin/config/ai/similar-content
Configuration options include:
- Enable/disable similarity search
- Set the similarity threshold (e.g., 70%)
- Select which content types to scan for similarity
- Adjust the debounce delay for suggestions (e.g., 300ms)
After configuration, the module will automatically inject its functionality into text inputs, textareas, and supported WYSIWYG editors. Users will see suggestions appear below their inputs as they type.
Additional Requirements
To work properly, this module requires:
- Milvus Vector Database: Handles similarity search using vector embeddings.
- AI Provider Plugin: Generates embeddings from input text. You can plug in your own embedding provider if needed.
Recommended modules/libraries
- CKEditor (4 or 5): For WYSIWYG support
- Milvus Integration Module: Facilitates connection between Drupal and Milvus
Similar projects
- Search API: Offers flexible full-text and faceted search capabilities, but is not focused on real-time semantic similarity.
- Entity Reference Autocomplete: Helps relate entities, but relies on title matches and lacks AI-driven similarity features.
What differentiates this module:
AI Similar Content is the only real-time module leveraging embeddings and vector search to return semantically similar results as the user types.
Supporting this Module
Want to support development? Consider:
- Contributing on GitHub
- Sponsoring via OpenCollective, GitHub Sponsors, or other channels
Community Documentation
- YouTube Walkthrough: Watch the setup and usage demo
- Demo Site: Explore it live in action
- External Docs: In-depth documentation on configuration and customization options