04-langchainjs

Demo - Speed up your AI development with LangChain.js

Prerequisites

• Visual Studio Code

All prerequisites from this sample are required, so you can run it both locally and using Azure OpenAI.

Scenario

Preparation

1. Clone the sample repo on your machine: https://github.com/Azure-Samples/langchainjs-quickstart-demo
2. Open the code in VS Code
3. Create the Azure resources needed for the sample, following the instructions in the README file, and fill in the .env file with the required information.
4. Open the local.js file in the editor.

Demo

1. Open the code from this repo in VS Code: https://github.com/Azure-Samples/langchainjs-quickstart-demo
2. Open local.js, and explain the code step by step.
3. Run the demo using play button
4. Update the prototype to use Azure:

• Replace model/db imports (use the imp snippet)
• Replace the model init section (use Copilot or newai snippet for the AI Search part
• Explain that the models are defined in .env file
• Replace the embeddings part
• Use the sea snippet to first need to check if documents are already indexed
• Use add snippet (or use Copilot) to complete the embedding part
• Run the demo again using play button

We did not have to change any of the workflow logic, as you’ve seen the migration is quite straightforward. Now I can take this code and wrap it in a serverless function to deploy it to Azure.