Semantic and Contextual Search
Unbody's Semantic and Contextual Search feature combines the power of Large Language Models (LLMs) and Weaviate, a potent vector database, to deliver multimodal semantic search capabilities. This feature transcends the boundaries of traditional keyword-based searches and understands the context and the intent of your queries, offering an unparalleled content discovery experience.
Under the hood, Unbody's Semantic and Contextual Search leverages the advanced technologies of "Retrievers & Vectorizers" modules provided by Weaviate. These modules, which include text2vec and multi2vec, are capable of processing diverse types of content including text and multimedia. Particularly, the text2vec module utilizes two vectorizers:
text2vec-contextionary. These vectorizers allow for nuanced, context-aware searching across various types of content.
You can choose between the
openai text vectorizers through the Unbody Dashboard. Here's how to set it up:
- Log into your Unbody Dashboard.
- Navigate to the 'Settings' tab.
- In the 'Search' section, you'll find the 'Vectorizer' dropdown menu.
- Choose either
openaias per your requirement.
contextionary vectorizer is a fast and efficient vectorizer that performs well for most standard use cases. If you need to handle more complex, nuanced text data, the
openai vectorizer, powered by Large Language Models, can provide more accurate and contextually aware results.
- Content Management Systems: Unbody's Semantic and Contextual Search can be implemented in content management systems, improving search efficiency by providing precise results based on the semantic and contextual relevance of the query.
- Customer Support Platforms: Implement this feature in customer support platforms to empower support agents with the ability to find relevant solutions by searching with natural language queries.
To use the Semantic and Contextual Search feature, ensure that Unbody is correctly set up and integrated within your application. Detailed setup instructions can be found in the Getting Started Guide.
Post setup, you can use the Semantic and Contextual Search feature by making API calls. Here's a sample:
//Sample API call //Replace 'your_query' with your search term