LangChain4J demo

Author: Julien Dubois

Video

This demo is supporting a video called "LangChain4J - use the power of LLMs in Java!", available at https://youtu.be/x8kkjmCZTaw.

Goal

This is a Spring Boot project that demonstrates how to use LangChain4J to create Java applications using LLMs.

It contains 7 demos:

• How to generate an image using Dalle-3.
• How to generate a text using GPT-4 and Mistral 7B.
• How to use a chat conversation with memory of the context.
• How to ingest data into a vector database, and use it.

Configuration

Azure OpenAI

You need to set the following environment variables:

• AZURE_OPENAI_ENDPOINT: your Azure OpenAI URL endpoint.
• AZURE_OPENAI_KEY: your Azure OpenAI API key.

MistralAI

You can use a local version of Mistral 7B, and you need to set following environment variables:

• MISTRAL_AI_BASE_URL: your MistralAI URL endpoint (for example http://localhost:1234/v1/ when running locally).
• MISTRAL_AI_KEY: your MistralAI API key (which you would not use when running the model locally, so you can use foo as a value).

Azure AI Search

You need to set the following environment variables:

• AZURE_SEARCH_ENDPOINT: your Azure AI Search URL endpoint.
• AZURE_SEARCH_KEY: your Azure AI Search API key.

Running the demos

You can run the demos using the following command:

./mvnw spring-boot:run

Then you can access the base URL, where you find the Web UI:

http://localhost:8080/