MemoryGraph – Git for knowledge, built to capture research dead ends

Short summary

MemoryGraph is an open-source knowledge graph tool that captures research workflows—observations, hypotheses, conclusions, dead ends, and open questions—as typed, fully-versioned nodes. It addresses the $100B annual research duplication problem by replacing static documentation with git-like collaboration: commits, forks, PRs, and diffs. Built on Kuzu, LLM-agnostic, AGPL-3.0, actively seeking contributors and builders.

• •Captures research dead-ends and failed hypotheses as first-class nodes in a knowledge graph with full version history
• •Git-inspired collaboration: can fork subgraphs, propose merges with conflict detection, compute semantic deltas between research trajectories
• •Open-source (AGPL-3.0), embedded Kuzu DB, works with any LLM via structured prompting; GitHub link provided with clear contribution paths

Generated with AI, which can make mistakes.