Trovares becomes Rocketgraph, launches AI-driven front end

We last covered Trovares in 2019, when it had only just come out of stealth mode. At the time, we were impressed by the performance of its in-memory property graph engine, its capacity for query processing on extremely large graphs, and its adoption by large companies and (we strongly suspected) government agencies. Six years later, it has re-emerged as Rocketgraph, retaining its highly performant and secure back end while adding a user-friendly front end, complete with graph visualisation and the incorporation of generative AI.

The aforementioned front end – Rocketgraph “Mission Control” – provides a conventional, graphical user interface for leveraging the company’s graph engine. Crucially, this makes the product accessible for all kinds of users. Moreover, Mission Control can integrate with your on-site LLMs using a pluggable interface. Once you have loaded your data into memory, these LLMs can be leveraged to generate an appropriate graph schema for that data, along with an explanation for the decisions made during that generation. The schema and data can then be combined to automatically create a corresponding (and visualised) graph.

This is not the only element of the platform driven by generative AI. It also includes “Rocketchat”, an AI-powered chatbot that can be queried for information and advice. Notably, you can include schemas with your queries in order to ask about them more specifically. Data, of course, is never included. Rocketgraph also uses generative AI to convert natural language query descriptions into executable Cypher code. As with schema generation, there is some explainability here: AI is used to create a summary of what the generated code does, as well as why and when you might want to use it. The latter in particular is designed to help newcomers understand and learn Cypher. The platform also provides some validation for these queries, in the form of syntax checking, with automatic regeneration of the query if the check is failed.

When taken together, these additions mark a shift from a strictly back end tool, when the company was still Trovares, to a much more accessible graph analytics platform, now that it has rebranded to Rocketgraph. We have always been impressed by the engine’s performance, scalability, and security (qualities that have certainly been maintained), so we are very happy to see it enhancing its usability and thereby broadening its overall appeal.