{"slug": "i-reverse-engineered-the-three-biggest-agent-memory-tools", "title": "I reverse-engineered the three biggest agent-memory tools", "summary": "A developer reverse-engineered three major agent-memory tools—Cognee, Graphiti, and Neo4j's agent-memory—finding they all rely on heavy knowledge-graph architectures. The author argues such infrastructure is overkill for personal or small-scale use, preferring simpler setups like Obsidian and LLM wikis, while acknowledging the value of graph-based tools for medium-to-large scale products.", "body_md": "I spent weeks reading about how Cognee, Graphiti, and Neo4j's `agent-memory` build their agent memory architectures. They converged on the same heavy knowledge-graph design: an ontology, LLM extraction pipelines, deduplication, the works.\n\nI really wanted to use them for my personal use case, but that looks like such a heavy setup that adds a lot of friction and silos. Plus, it feels like I just get my data trapped in their service, for not a ton of value.\n\nThat's why my \"long-term memory\" still lives in Obsidian, Readwise, and Google Drive, with per-project LLM wikis as the agent's memory. No infrastructure. And I'm fine with it.\n\nThey ship memory as a product, which, in my opinion, at a personal or small scale, is overkill. You can build the same \"knowledge graph\" experience via plain old `.md` files within an LLM wiki memory.\n\nBut still, graphs are strong, so I adapted the same architecture from the Cognee, Graphiti, and Neo4j `agent-memory` stacks to build a data-mining tool with just MongoDB, VoyageAI, and Gemini Flash. But I scoped it to a very particular problem and ontology domain to avoid the KG noise.\n\nOn the other end of the spectrum, if you want to ship a product at medium-to-large scale, it makes sense to start using monsters such as Neo4j, Zep, or HydraDB.\n\nBut I am curious: what is your long-term memory setup? Obsidian + LLM wikis vs. Cognee/Graphiti/Zep? Do you actually use tools such as Cognee or Zep?\n\nIn case you are curious about how Cognee, Graphiti, and Neo4j's `agent-memory` work under the hood, I wrote a full breakdown here: https://www.decodingai.com/p/unified-memory-from-scratch-knowledge-graphs\n\nComments URL: [https://news.ycombinator.com/item?id=48919162](https://news.ycombinator.com/item?id=48919162)\n\nPoints: 1\n\n# Comments: 0", "url": "https://wpnews.pro/news/i-reverse-engineered-the-three-biggest-agent-memory-tools", "canonical_source": "https://news.ycombinator.com/item?id=48919162", "published_at": "2026-07-15 11:23:02+00:00", "updated_at": "2026-07-15 11:47:52.899850+00:00", "lang": "en", "topics": ["ai-agents", "ai-tools", "ai-infrastructure", "developer-tools"], "entities": ["Cognee", "Graphiti", "Neo4j", "Obsidian", "Readwise", "Google Drive", "VoyageAI", "Gemini Flash"], "alternates": {"html": "https://wpnews.pro/news/i-reverse-engineered-the-three-biggest-agent-memory-tools", "markdown": "https://wpnews.pro/news/i-reverse-engineered-the-three-biggest-agent-memory-tools.md", "text": "https://wpnews.pro/news/i-reverse-engineered-the-three-biggest-agent-memory-tools.txt", "jsonld": "https://wpnews.pro/news/i-reverse-engineered-the-three-biggest-agent-memory-tools.jsonld"}}