Tutorial: Indexing text with OpenAI embeddings

This tutorial shows how to build a small Hacker News search tool with Vecal and OpenAI embeddings. The complete code lives in src/main.ts of the project.

1. Obtain an OpenAI API key.
2. Fetch the data you want to index.
3. Convert each text item to an embedding using the OpenAI API.
4. Create a VectorDB instance with the correct dimension and store each vector.
5. Perform searches with db.search or db.annSearch.

const db = new VectorDB({ dbName: 'hn-stories', dimension });

await db.add(embedding, { title: story.title, url: story.url });
const results = await db.search(queryEmbedding, 5);

Check the source file for the full working example.