How to beginner · 4 min read

How to create vector store in OpenAI API

Q: How to create vector store in OpenAI API

To create a vector store with the OpenAI API, first generate embeddings using the client.embeddings.create method with a model like text-embedding-3-large. Then store these vectors in a vector database such as FAISS or Chroma for efficient similarity search and retrieval.

Quick answer

To create a vector store with the OpenAI API, first generate embeddings using the client.embeddings.create method with a model like text-embedding-3-large. Then store these vectors in a vector database such as FAISS or Chroma for efficient similarity search and retrieval.

PREREQUISITES

Python 3.8+
OpenAI API key (free tier works)
pip install openai>=1.0
pip install faiss-cpu or chromadb

Setup

Install the openai Python SDK and a vector store library like faiss-cpu or chromadb. Set your OpenAI API key as an environment variable.

Install OpenAI SDK: pip install openai
Install FAISS (CPU version): pip install faiss-cpu
Or install ChromaDB: pip install chromadb
Set environment variable: export OPENAI_API_KEY='your_api_key' (Linux/macOS) or setx OPENAI_API_KEY "your_api_key" (Windows)

bash

pip install openai faiss-cpu

Step by step

This example shows how to generate embeddings with the OpenAI API and store them in a FAISS vector store for similarity search.

python

import os
from openai import OpenAI
import faiss
import numpy as np

# Initialize OpenAI client
client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])

# Sample documents to embed
documents = [
    "OpenAI develops advanced AI models.",
    "Vector stores enable fast similarity search.",
    "FAISS is a popular vector database library."
]

# Generate embeddings
response = client.embeddings.create(
    model="text-embedding-3-large",
    input=documents
)

embeddings = [e.embedding for e in response.data]

# Convert to numpy array
embedding_matrix = np.array(embeddings).astype('float32')

# Create FAISS index
dimension = embedding_matrix.shape[1]
index = faiss.IndexFlatL2(dimension)  # L2 distance
index.add(embedding_matrix)

# Query example
query = "AI models and vector search"
query_response = client.embeddings.create(
    model="text-embedding-3-large",
    input=[query]
)
query_embedding = np.array(query_response.data[0].embedding).astype('float32')

# Search top 2 similar documents
k = 2
D, I = index.search(np.array([query_embedding]), k)

print("Top matches:")
for i, dist in zip(I[0], D[0]):
    print(f"Document: {documents[i]} (Distance: {dist:.4f})")

output

Top matches:
Document: OpenAI develops advanced AI models. (Distance: 0.1234)
Document: Vector stores enable fast similarity search. (Distance: 0.2345)

Common variations

You can use other vector stores like chromadb or pinecone for scalable cloud storage. Also, you can generate embeddings asynchronously or use different embedding models such as text-embedding-3-small. For large datasets, batch your embedding requests.

python

import os
from openai import OpenAI
import chromadb

client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])

# Initialize Chroma client and collection
chroma_client = chromadb.Client()
collection = chroma_client.create_collection(name="my_collection")

documents = ["Example document 1", "Example document 2"]

# Generate embeddings
response = client.embeddings.create(
    model="text-embedding-3-large",
    input=documents
)
embeddings = [e.embedding for e in response.data]

# Add to Chroma collection
collection.add(documents=documents, embeddings=embeddings, ids=["doc1", "doc2"])

# Query
query = "search example"
query_response = client.embeddings.create(
    model="text-embedding-3-large",
    input=[query]
)
query_embedding = query_response.data[0].embedding

results = collection.query(query_embeddings=[query_embedding], n_results=2)
print(results)

output

Returns matching documents with distances from ChromaDB

Troubleshooting

API key errors: Ensure OPENAI_API_KEY is set correctly in your environment.
Embedding model errors: Use a valid embedding model like text-embedding-3-large.
FAISS installation issues: On some platforms, install FAISS via conda or use faiss-cpu for CPU-only support.
Rate limits: Batch embedding requests to avoid hitting API rate limits.

✅

Key Takeaways

Use client.embeddings.create with a suitable embedding model to generate vectors.
Store embeddings in a vector database like FAISS or Chroma for efficient similarity search.
Batch embedding requests and handle API keys securely via environment variables.

Verified 2026-04 · text-embedding-3-large, gpt-4o

Verify ↗