Show HN: Genesys – Causal graph memory for AI agents, not just vectors

Category: infrastructure

Tags: ai-memory, causal-graph, mcp, agent-framework, knowledge-graph, memory-management

Score: 7.5/10 (Innovation: 7, Technical: 8, Documentation: 8, Utility: 7)

Genesys is a causal graph memory system for AI agents that scores, links, and prunes memories using a multiplicative decay formula, moving beyond flat vector stores. It integrates natively with the Model Context Protocol (MCP) and supports multiple storage backends including Postgres, Obsidian, and FalkorDB. Its combination of causal reasoning, lifecycle management, and strong benchmark results makes it a promising solution for persistent AI agent memory.

Target audience: AI engineers, agent developers, backend devs

Repository: https://github.com/rishimeka/genesys · Python · AGPL-3.0 · 12 stars

View on Hacker News