LangChain.js + Azure Cosmos DB RAG Quickstart

Get started with Retrieval-Augmented Generation (RAG) and vector search using LangChain.js and Azure Cosmos DB.

What this sample demonstrates

1. Document ingestion — Load text, split into chunks, and store vector embeddings in Azure Cosmos DB
2. Semantic similarity search — Find relevant documents using vector similarity
3. Score-threshold search — Filter results by similarity score
4. MMR (Maximal Marginal Relevance) — Retrieve diverse, non-redundant results
5. RAG chain — Answer questions using retrieved context with Azure OpenAI GPT-4.1

Prerequisites

• Node.js v18 or later
• An Azure Cosmos DB account with NoSQL API
• An Azure OpenAI deployment with:
  • A chat model (e.g., gpt-4.1)
  • An embedding model (e.g., text-embedding-3-small)

Getting started

1. Clone the repo

   git clone https://github.com/AzureCosmosDB/langchainjs-cosmosdb-rag-quickstart.git
   cd langchainjs-cosmosdb-rag-quickstart

2. Install dependencies

   npm install

3. Configure environment variables

   cp .env.example .env

   Edit .env and fill in your Azure credentials.

4. Run the sample

   npm start

Project structure

File	Description
get-started.js	Main application — ingestion, search, and RAG
sample-data.txt	Sample document about Azure Cosmos DB
.env.example	Template for required environment variables

Resources

• Azure Cosmos DB Vector Search
• LangChain.js Azure Cosmos DB Integration
• Azure OpenAI Service