cosmic-community/semantic-search-example

This project is a sample documentation website that shows off a feature called semantic search from a headless content management service called Cosmic. Instead of matching exact keywords the way traditional search boxes do, semantic search understands the meaning behind a question and returns documentation articles that are actually relevant, even if the words do not match exactly. The site itself is a clean, modern docs library where articles are organized into categories and document types, with pages you can browse and filter. It was built with Next.js, a popular framework for building fast, search engine friendly websites, along with TypeScript and Tailwind CSS for styling. The pages are rendered on the server for speed, and the design is fully responsive, meaning it works well on phones, tablets, and desktop screens alike. The repository was generated automatically by Cosmic's AI tools based on two simple prompts: one describing the content structure needed and one describing the site itself. Because of this, the project doubles as an example of what Cosmic's AI code generation can produce from plain instructions, and it is set up to be cloned directly into your own Cosmic account along with its matching content bucket. To run it yourself, you need a Cosmic account with two types of content already set up, categories and docs, plus a bucket slug and API keys, which get placed into environment variables. Semantic search specifically requires a write key rather than just a read key, since it calls Cosmic's Content RAG search endpoint. Once installed, the project runs locally with a single development server command, and it can also be deployed to hosting services such as Vercel or Netlify with the same environment variables configured there.