A variety of open-source and community-supported frameworks are available to help you integrate Neo4j with generative AI and large language models (LLMs).
These frameworks support use cases such as:
-
Retrieval-augmented generation (RAG).
-
Agentic workflows.
-
Knowledge graph construction.
Key Features of GenAI Frameworks
Typically these frameworks include:
-
LLM usage, prompt and output management
-
Embedding model integration
-
Vector and database integration (including Neo4j)
-
RAG (Retrieval-Augmented Generation) workflows
-
Agentic workflows and orchestration
-
Monitoring, observability, and deployment tools
Popular GenAI Frameworks for Neo4j
-
LangChain (Python) - A leading open-source framework for building LLM-powered applications, with strong support for Neo4j as a vector store and knowledge graph.
-
LangChainJS - The JavaScript/TypeScript version of LangChain, enabling GenAI workflows in Node.js and browser environments.
-
LlamaIndex- A data framework for connecting LLMs to external data, with connectors for Neo4j to support RAG and knowledge graph use cases.
-
Spring AI - A Spring ecosystem project for integrating AI capabilities into Java applications, including Neo4j support.
-
LangChain4j - A Java implementation of LangChain, supporting Neo4j integration for LLM and RAG workflows.
-
Haystack - An open-source framework for building search and question-answering systems, with Neo4j integration for graph-based retrieval.
-
Semantic Kernel - A Microsoft open-source orchestration library for AI workflows, supporting Neo4j as a data source.
-
DSPy - A framework for programming and optimizing LLM pipelines, with Neo4j connectors.
You can learn more in the Neo4j GenAI Frameworks documentation.
Frameworks support your GenAI application development and help you build complex workflows that integrate Neo4j with LLMs and other AI components.
Check Your Understanding
Features of GenAI Frameworks
Which of the following are typical features of GenAI frameworks? (Select all that apply)
-
✓ Embedding model integration
-
✓ RAG (Retrieval-Augmented Generation) workflows
-
✓ Agentic workflows and orchestration
-
❏ Data storage and query language
Hint
GenAI frameworks are designed to help you manage LLMs, embeddings, and workflows, and integrate with other technologies such as LLMs and databases.
Solution
The correct answers are:
-
Embedding model integration
-
RAG (Retrieval-Augmented Generation) workflows
-
Agentic workflows and orchestration
GenAI frameworks are designed to help you manage LLMs, embeddings, and workflows, and integrate with other technologies such as LLMs and databases. They do not typically provide their own data storage or query language.
Lesson Summary
In this lesson, you learned about some of the popular GenAI frameworks that integrate with Neo4j, including LangChain, LlamaIndex, and Spring AI.
In the last lesson, you will review your progress and explore some of the next steps to continue your learning.