Memgraph 3 has been updated with the addition of two new tools; an AI Graph Toolkit that automates the porting of data into a knowledge graph in Memgraph; and an MCP Client within Memgraph Lab. 

The open source Memgraph graph database is compatible with Neo4j, is ACID compliant and uses the Cypher query language for structuring, manipulating and exploring data. Memgraph is built in C/C++ and uses an in-memory first architecture. Users can run Python, Rust, and C/C++ code natively, and can ingest data from sources including Kafka, SQL and CSV files.

Memgraph features include support for deep-path traversals, and the ability to use advanced capabilities such as accumulators and path filtering without adding additional application logic. 

The toolkit has been developed to make it easier to get SQL and unstructured data into a graph format ready to be used by various search techniques and algorithms within a GraphRAG pipeline. For SQL data, this required mapping tables, identifying entities, and performing entity resolution to "graphify" the structures. For unstructured data, it involved chunking, cleaning, and creating vector embeddings, to reach the natural language/ChatGPT interfaces wanted.

The new toolkit and MCP client allow engineering teams to bypass the manual coding, programming, and data. The Memgraph team says engineers will still need to fine-tune the final output, but the tedious work of extracting and transforming data from SQL and unstructured formats is already taken care of.

SQL2Graph is powered by Memgraph's Hypothetical Graph Model (HyGM). It analyzes your SQL schema, proposes an initial graph model based on it, then the user can iteratively refine it before bringing the data into Memgraph. The tool can be used in two ways; and automatic mode for straightforward schemas, and a table-by-table review with interactive refinement for complex databases.

Memgraph describes HyGM as effectively HyDE for knowledge graphs, but instead of generating hypothetical documents, it generates hypothetical graph models.
HyGM  hypothesizes an initial graph model from your data, and explains its reasoning in human terms. It then refines the model based on your input, and validates to make sure the result makes sense.

Memgraph's Chief Technology Officer and Co-founder, Marko Budiselić, says:

"The AI Graph Toolkit makes it extremely easy for developers to transform SQL and unstructured data into knowledge graphs, and then produce GraphRAG-level data structures to superpower your AI chatbot capabilities."

Budiselić says the toolkit means users can now run GraphRAG against the ideal back-end LLM input of relational data tables with numeric values and all the business context currently trapped in unstructured text, PDF, and document forms.

"Engineering teams will be able to much more easily unleash the power of graph and GraphRAG across multiple forms of business data, enabling chatbot natural language access and querying capabilities to the whole corporate back end."

The new Memgraph AI Toolkit is available for download through Github and can be used via Memgraph Cloud.

Memgraph has also launched a JumpStart Programme. This is a package that pairs a Memgraph Enterprise license with 20+ hours of hands-on help and implementation, powered by the Memgraph AI Toolkit.

The Toolkit will be complemented by the MCP client, available later this month. This is designed to make it easy to connect Memgraph's graph capabilities to multiple back-end data sources through other MCP Servers, while supporting adoption of an industry standard to accelerate developer AI productivity. 

Memgraph 3.0 is available for download now.

More Information

Memgraph Download Page

Memgraph On GitHub

Memgraph Website

Related Articles

Memgraph 3 Simplifies Graph Based AI Projects

To be informed about new articles on I Programmer, sign up for our weekly newsletter, subscribe to the RSS feed and follow us on Twitter, Facebook or Linkedin.

Comments

Make a Comment or View Existing Comments Using Disqus

or email your comment to: comments@i-programmer.info