Chroma: Getting Started
What you'll learn: How to install Chroma and perform your first vector operations using both in-memory and persistent storage modes.
What is Chroma?
Chroma is an open-source vector database designed specifically for AI applications. Think of it as a specialized filing system that organizes information by meaning rather than keywords — perfect for storing and searching embeddings you generate from text, images, or other data.
Installation
Chroma installs easily via pip:
pip install chromadb
Creating Your First Collection
A collection in Chroma is like a table in a traditional database — it holds your vectors and their metadata:
import chromadb
# In-memory mode (data lost when program stops)
client = chromadb.Client()
# Create a collection
collection = client.create_collection(name="my_documents")
Basic Vector Operations
Here's how to add vectors and search them:
# Add vectors with IDs, embeddings, and optional metadata
collection.add(
ids=["doc1", "doc2", "doc3"],
embeddings=[[0.1, 0.2, 0.3], [0.4, 0.5, 0.6], [0.7, 0.8, 0.9]],
metadatas=[{"source": "web"}, {"source": "book"}, {"source": "web"}]
)
# Query with a vector
results = collection.query(
query_embeddings=[[0.1, 0.2, 0.3]],
n_results=2
)
Persistent Storage
To keep your data between sessions:
client = chromadb.PersistentClient(path="./chroma_data")
collection = client.get_or_create_collection(name="my_documents")
Key Takeaway: Chroma provides a lightweight, Python-native way to store and query vectors with just a few lines of code, supporting both temporary in-memory experiments and persistent production storage.