Integrating LangChain With Azure ML Prompt Flow — Part 8

2 min readFeb 15, 2024

This post is in continuation to my previous articles, where I explained about Azure Machine Learning Prompt Flow. Here are those:

Part 1 — Getting started with Azure Prompt Flow

Part 2 — Create, update and test Azure Prompt Flow locally

Part 3 — Chat With Custom Data — FAISS Index & Azure ML Prompt Flow

Part 4 — Integrating Azure AI Search with Azure Prompt Flow

Part 5 — Use multiple prompts to analyze LLM response — Prompt Variant

Part 6 — Deploy And Consume LLM App Using Azure ML Prompt Flow

Part 7 — Evaluate Azure ML Prompt Flow Using Built-in Methods

Till now, I’ve covered the Azure Prompt Flow basics wherein we didn’t pull anything from outside. Let’s move a step further and see, how can we customize our flow using LangChain.

Think of a scenario, wherein you have already written a piece of code utilizing LangChain and is very well tested. Now, you want to integrate that code to Prompt Flow.

Integrate LangChain With Prompt Flow

In order to integrate LangChain, either we can create customized environment having all the dependencies installed or we can create runtime based on our local settings.

In below video, I have explained the simplest possible way to integrate LangChain by utilizing existing tool/node. Here is the snippet for that: