Inspiration
AI agents are increasingly becoming the worlds most trending tools. They are becoming part of our daily lives. However with rise of AI agents come with the rise of Knowledge Graphs and Context Graphs as a supporting tool for AI agents acting as fact storage and brain for AI agents respectively. This bring the need to have graph databases that are natively built for AI. A database that can easily scale at large store many graphs in isolation and be really fast in query retrieval as the graph grows in size. Hence why i built Graphmind.
What it does
Graphmind is a high-performance, distributed graph database written in Rust with 100% OpenCypher support, Redis protocol compatibility, and a built-in web visualizer. It combines property graph storage, vector search, graph algorithms, multi tenancy, and natural language querying in a single binary. All this are building block for a support for AI agents in
How we built it
Graphmind is built with Rust, A grammar was developed for it, which then went through the parser and the planner. More on that API was developed to interface with Graphmind over various programming languages like python and rust. I added multi tenancy to allow scaling for projects within the same environment and for easy use with AI.
Challenges we ran into
Converting cypher grammar into pest which is rust compatible is very hard to do. However i was able to do that carefully and passed even more cypher syntax that is dedicated to Graphmind database
Accomplishments that we're proud of
I proud to have been able to build a property graph database from scratch one that i can now use for my projects free from any licenses and limitations.
What we learned
What's next for Graphmind
Adding Integration with AI packages like graphmind-context-graphs that will act as brain for AI.
Log in or sign up for Devpost to join the conversation.