Welcome to the Healthsearch Demo, an open-source project that showcases how user-written reviews and queries can be used to retrieve supplement products based on specific health effects. The search functionality in this demo accepts natural language queries and translates them into GraphQL queries using Large Language Models (LLMs) like GPT4. These GraphQL queries are then used to retrieve supplements from a Weaviate database. The demo also features generative search, where product summaries are generated based on the retrieved objects. It’s important to note that the results and summaries should not be considered health advice, as they are solely based on user-written reviews. Healthsearch relies on semantic search within user reviews to find products that are good for specific health conditions. The demo scans user reviews for discussions on products that have alleviated similar conditions and aggregates the results. Generative Search is another feature of the demo that uses an LLM to generate concise summaries of the top five results. These summaries highlight the pros and cons of the products and provide valuable insights. The demo also utilizes a Semantic Cache by embedding the generated results and queries to Weaviate. This allows the demo to return results from queries that are semantically similar to new queries, improving the search process. The Healthsearch Demo serves as
https://github.com/weaviate/healthsearch-demo