Chroma
Embedded vector database — simplest path to production RAG
📋 Overview
Chroma is an embedded vector database designed for the simplest path from prototype to production RAG. Unlike Qdrant (which requires a server), Chroma runs embedded in your Python process — no separate server needed. It provides a simple API for adding documents, querying by similarity, and filtering by metadata.
✨ Key Features
- •22K+ GitHub stars — simplest vector database
- •Embedded mode — no separate server needed
- •Simple API: add documents, query, done
- •Persists to disk by default
- •Server mode for production
- •Apache 2.0 licensed
- •Simple API
- •Local persistence
🎯 The Problem It Solves
RAG applications need a vector database that is trivial to set up and use. Chroma provides the simplest API of any vector DB.
🔧 How It Works
Chroma runs embedded in your Python process — call chroma_client.create_collection(), add documents with embeddings, and query with collection.query(). It persists to disk by default and can run as a server for production.
🚀 Installation & Quick Start
Installation
See websiteQuick Start
- See documentation
✅ Pros
- •Simplest vector database API
- •Embedded mode for prototyping
- •Apache 2.0 licensed
- •Simple API
- •Local persistence
- •Active community
❌ Cons
- •Limited scalability
- •No distributed mode
- •Server mode less mature
- •Limited scalability
💬 Practitioner Verdict
“Chroma is the simplest vector database — the embedded mode is genuinely differentiated for prototyping. The trade-off: limited scalability compared to Qdrant, no distributed mode, and the server mode is less mature.”
Self-Hosted (Free)
Open source, MIT/Apache licensed. Run it yourself.
⭐ Star & Clone on GitHubFree forever. Your infrastructure, your data.
📊 Specifications
- Language
- Python/Rust
- License
- Apache-2.0
- Platform
- Linux, macOS, Windows
- Supported Models
- REST API, CLI
💰 Pricing Reality
100% free Apache 2.0. No paid tier.