DEV Community

Alex Spinov
Alex Spinov

Posted on

ChromaDB Has a Free API — Heres How to Add Embeddings to Any Python App

#ai #python #database #tutorial

ChromaDB is an open-source embedding database — the simplest way to add AI-powered search, RAG, and recommendations to your Python or JavaScript app.

Why ChromaDB?

  • Simplest API: 4 functions to learn
  • Auto-embedding: Built-in embedding models
  • Python + JS: Both SDKs available
  • Self-hosted: No cloud dependency
  • Metadata filtering: Combine semantic + keyword search
  • Persistent storage: SQLite backend

Install 

pip install chromadb
Enter fullscreen mode Exit fullscreen mode

Basic Usage 

import chromadb

client = chromadb.Client()

# Create collection (auto-embeds with default model)
collection = client.create_collection('articles')

# Add documents — ChromaDB embeds them automatically
collection.add(
    documents=[
        'Introduction to machine learning with Python',
        'Building REST APIs with FastAPI',
        'Docker containers for beginners',
        'React hooks explained simply',
    ],
    ids=['doc1', 'doc2', 'doc3', 'doc4'],
    metadatas=[
        {'category': 'AI', 'level': 'beginner'},
        {'category': 'Backend', 'level': 'intermediate'},
        {'category': 'DevOps', 'level': 'beginner'},
        {'category': 'Frontend', 'level': 'intermediate'},
    ]
)

# Search
results = collection.query(
    query_texts=['How do I build an API?'],
    n_results=2
)
print(results['documents'])
# [['Building REST APIs with FastAPI', 'React hooks explained simply']]
Enter fullscreen mode Exit fullscreen mode

Persistent Storage 

client = chromadb.PersistentClient(path='./chroma-data')
Enter fullscreen mode Exit fullscreen mode

Server Mode 

# Start server
chroma run --host 0.0.0.0 --port 8000
Enter fullscreen mode Exit fullscreen mode 
# Connect from client
client = chromadb.HttpClient(host='localhost', port=8000)
Enter fullscreen mode Exit fullscreen mode

Filtered Search 

results = collection.query(
    query_texts=['backend tutorial'],
    n_results=5,
    where={'category': 'Backend'},
    where_document={'$contains': 'API'}
)
Enter fullscreen mode Exit fullscreen mode

With Custom Embeddings (OpenAI) 

from chromadb.utils import embedding_functions

openai_ef = embedding_functions.OpenAIEmbeddingFunction(
    api_key='sk-...',
    model_name='text-embedding-3-small'
)

collection = client.create_collection(
    name='articles',
    embedding_function=openai_ef
)
Enter fullscreen mode Exit fullscreen mode

RAG in 10 Lines 

import chromadb
import openai

db = chromadb.PersistentClient(path='./docs')
collection = db.get_collection('knowledge-base')

def ask(question: str) -> str:
    docs = collection.query(query_texts=[question], n_results=3)
    context = '\n'.join(docs['documents'][0])
    return openai.chat.completions.create(
        model='gpt-4',
        messages=[{'role': 'user', 'content': f'Based on: {context}\n\nAnswer: {question}'}]
    ).choices[0].message.content
Enter fullscreen mode Exit fullscreen mode

Update and Delete 

# Update
collection.update(
    ids=['doc1'],
    documents=['Updated: Advanced ML with Python and scikit-learn'],
    metadatas=[{'category': 'AI', 'level': 'advanced'}]
)

# Delete
collection.delete(ids=['doc4'])

# Delete by filter
collection.delete(where={'category': 'Frontend'})
Enter fullscreen mode Exit fullscreen mode

Real-World Use Case

A support team embedded 5,000 help articles into ChromaDB. When customers typed questions, the system found relevant articles in <50ms — reducing ticket volume by 40% because customers found answers themselves.

Need to automate data collection? Check out my Apify actors for ready-made scrapers, or email spinov001@gmail.com for custom solutions.

Top comments (0)