Weaviate — Animate | AI Wisdom

← Back to topic
🎞 ANIMATE · Mode 3 of 5 · +100 XP— Weaviate Vector Database
📋Connect
⚙️Schema
🔄Ingest
✅Hybrid
🎯 Next: Test →
Step 1 / 4
Speed
Connect & Configure
weaviate.client with OpenAI API key — Weaviate calls embedding API on your behalf💡 Analogy
Like a library with its own cataloging staff — you hand over books, they create the index cards (embeddings).
Weaviate connects to OpenAI for automatic vectorization — you send text, Weaviate embeds it. Pass your OpenAI key in the header so Weaviate can make embedding calls.
Console Output
> Connected to Weaviate: https://my-cluster.weaviate.network
> OpenAI API key configured for text2vec-openai
> Modules: text2vec-openai, hybrid-search
●weaviate.ts
Connect & ConfigureL1–9
1import weaviate, { WeaviateClient } from 'weaviate-ts-client';
2 
3// Connect to Weaviate instance
4const client: WeaviateClient = weaviate.client({
5  scheme: 'https',
6  host: process.env.WEAVIATE_HOST!,
7  apiKey: new weaviate.ApiKey(process.env.WEAVIATE_API_KEY!),
8  headers: { 'X-OpenAI-Api-Key': process.env.OPENAI_API_KEY! },
9});
10 
11// Define schema — Weaviate auto-embeds with text2vec-openai
12await client.schema.classCreator().withClass({
13  class: 'Article',
14  vectorizer: 'text2vec-openai',
15  moduleConfig: {
16    'text2vec-openai': { model: 'text-embedding-3-small' }
17  },
18  properties: [
19    { name: 'title', dataType: ['text'] },
20    { name: 'content', dataType: ['text'] },
21    { name: 'category', dataType: ['text'], tokenization: 'keyword' },
22  ],
23}).do();
24 
25// Import objects — Weaviate vectorizes automatically
26await client.batch.objectsBatcher()
27  .withObject({ class: 'Article', properties: {
28    title: 'Introduction to RAG', content: 'RAG retrieves...', category: 'AI' }})
29  .withObject({ class: 'Article', properties: {
30    title: 'Vector Databases', content: 'Vector DBs store...', category: 'DB' }})
31  .do();
32 
33// Hybrid search with Weaviate's built-in BM25 + vector
34const results = await client.graphql.get()
35  .withClassName('Article')
36  .withHybrid({ query: 'semantic vector search', alpha: 0.7 })
37  .withWhere({ path: ['category'], operator: 'Equal', valueText: 'AI' })
38  .withFields('title content _additional { score }')
39  .withLimit(5)
40  .do();
40 lines
Timeline — 4 Steps
Connect & Configure
weaviate.client with OpenAI API key — Weaviate calls embedding API on your behalf💡 Analogy
Like a library with its own cataloging staff — you hand over books, they create the index cards (embeddings).
Weaviate connects to OpenAI for automatic vectorization — you send text, Weaviate embeds it. Pass your OpenAI key in the header so Weaviate can make embedding calls.
Console Output
> Connected to Weaviate: https://my-cluster.weaviate.network
> OpenAI API key configured for text2vec-openai
> Modules: text2vec-openai, hybrid-search

Was this page helpful?

Discussion

weaviate.client with OpenAI API key — Weaviate calls embedding API on your behalf

💡 Analogy

Like a library with its own cataloging staff — you hand over books, they create the index cards (embeddings).

Weaviate connects to OpenAI for automatic vectorization — you send text, Weaviate embeds it. Pass your OpenAI key in the header so Weaviate can make embedding calls.

Console Output

> Connected to Weaviate: https://my-cluster.weaviate.network
> OpenAI API key configured for text2vec-openai
> Modules: text2vec-openai, hybrid-search

weaviate.client with OpenAI API key — Weaviate calls embedding API on your behalf

💡 Analogy

Like a library with its own cataloging staff — you hand over books, they create the index cards (embeddings).

Weaviate connects to OpenAI for automatic vectorization — you send text, Weaviate embeds it. Pass your OpenAI key in the header so Weaviate can make embedding calls.

Console Output

> Connected to Weaviate: https://my-cluster.weaviate.network
> OpenAI API key configured for text2vec-openai
> Modules: text2vec-openai, hybrid-search