How to beginner · 3 min read

How to store AI memory in vector database

Quick answer

Store AI memory by converting text or documents into vector embeddings using a model like text-embedding-3-small from OpenAI. Save these embeddings in a vector database such as FAISS or Chroma for efficient similarity search and retrieval during AI interactions.

PREREQUISITES

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

Setup

Install required Python packages and set your OpenAI API key as an environment variable.

Use pip install openai faiss-cpu chromadb to install dependencies.
Set OPENAI_API_KEY in your environment for authentication.

bash

pip install openai faiss-cpu chromadb

output

Collecting openai...
Collecting faiss-cpu...
Collecting chromadb...
Successfully installed openai faiss-cpu chromadb

Step by step

This example shows how to embed text memory with OpenAI embeddings and store them in a FAISS vector store for retrieval.

python

import os
from openai import OpenAI
from langchain_openai import OpenAIEmbeddings
from langchain_community.vectorstores import FAISS

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

# Sample AI memory texts
texts = [
    "The Eiffel Tower is in Paris.",
    "Python is a popular programming language.",
    "OpenAI develops advanced AI models."
]

# Create embeddings for texts
embedding_model = OpenAIEmbeddings(client=client, model="text-embedding-3-small")
vectors = embedding_model.embed_documents(texts)

# Create FAISS index and add vectors
index = FAISS.from_embeddings(vectors, texts)

# Query example: find closest memory to a query
query = "Where is the Eiffel Tower located?"
query_vector = embedding_model.embed_query(query)
results = index.similarity_search_by_vector(query_vector, k=1)

print("Query:", query)
print("Closest memory:", results[0].page_content)

output

Query: Where is the Eiffel Tower located?
Closest memory: The Eiffel Tower is in Paris.

Common variations

You can use other vector databases like Chroma or Weaviate for scalable storage. Async embedding calls are possible with OpenAI SDK. Different embedding models can be chosen based on accuracy and speed needs.

python

import os
import asyncio
from openai import OpenAI
from langchain_openai import OpenAIEmbeddings
from langchain_community.vectorstores import Chroma

async def async_embedding_example():
    client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])
    embedding_model = OpenAIEmbeddings(client=client, model="text-embedding-3-small")

    texts = ["AI memory storage in vector DB.", "Embedding vectors enable similarity search."]

    # Async embedding calls (simulate with asyncio.to_thread)
    vectors = await asyncio.to_thread(embedding_model.embed_documents, texts)

    # Store in Chroma vector DB
    chroma_index = Chroma.from_embeddings(vectors, texts)

    query = "How to store AI memory?"
    query_vector = await asyncio.to_thread(embedding_model.embed_query, query)
    results = chroma_index.similarity_search_by_vector(query_vector, k=1)

    print("Query:", query)
    print("Closest memory:", results[0].page_content)

asyncio.run(async_embedding_example())

output

Query: How to store AI memory?
Closest memory: AI memory storage in vector DB.

Troubleshooting

If embeddings fail, verify your OPENAI_API_KEY is set correctly.
If vector search returns irrelevant results, check that embeddings and queries use the same model.
For large datasets, ensure your vector database supports persistence and indexing.

✅

Key Takeaways

Use OpenAI embeddings like text-embedding-3-small to convert AI memory text into vectors.
Store embeddings in vector databases such as FAISS or Chroma for fast similarity search.
Ensure consistent embedding models for both storing and querying vectors to maintain accuracy.

Verified 2026-04 · text-embedding-3-small

Verify ↗