Zep AI (YC W24)

Knowledge graphs are huge for AI Agents! A knowledge graph is the difference between a dumb AI agent and one that blows everyone's mind. Agents need memory and must know how to keep it updated over time (This is difficult, and it's the main reason most agents you've seen get dumber overnight!) This is where a knowledge graph helps. A knowledge graph is a network of connected points, each representing a piece of information. It's a very efficient structure for capturing complex relationships between data. Google uses a (huge) knowledge graph as part of Search. (Probably the largest knowledge graph in the world.) It was arguably one of the best improvements to Search since it was created. For building AI agents, knowledge graphs have two advantages: 1. They make it easier to extract facts from memory 2. They make it easier to update facts as they change The second point is crucial: You want agents to keep up with the world and update old facts as they discover new information. Here is a recommendation that will teach you how to use a knowledge graph as the memory layer of an AI agent: Ken Collins wrote an excellent article in which he builds a chat history for Llama 3 using Zep AI (YC W24)'s AI Memory (backed by a knowledge graph.) Here is a link to the article: https://fnf.dev/4fPAXtx This article is a great example of how to build agents that keep up with change. Zep is an open-source library that will serve as your agent's memory. You can connect it to any agent framework, model, or platform. The article's source code is in TypeScript, but you can use Zep with Python or Go as well. In a few bullet points: 1. You send messages to your AI agent 2. Zep synthesizes the information into a knowledge graph 3. You can retrieve any relevant facts from memory extremely fast Thanks to the Zep team for sponsoring this post.

Zep Holiday Roundup: Explore User & Group Graphs, Fact Ratings, and SOC 2 Type 2 Certification! 🎄 🎁

Thanks for the shout out, Kesha Williams!

If you give a RAG agent chunks of space articles from 1994, it’ll tell you Pluto’s a planet. But what happens if you give it chunks of articles from 2004, and chunks from an article from 2024 that says Pluto is no longer a planet? The LLM might get confused, or bias towards the more frequent mention of Pluto being a planet. Vector databases have enabled AI agents to perform incredibly well at information retrieval. But RAG doesn’t do a great job of encoding time. Vector databases match for keyword and semantic relevance…not “How recent is this information?” or “Is this fact still valid?” Unless you design your system to account for this, you’ll get inaccurate results. That’s fine if time isn’t an issue for your AI agent. But if you’re designing an enterprise product that talks to users, it probably is: - Your customer support agent should know past issues your customers ran into - Your product recommendation agent should know the customer bought Nikes, then returned them because they “didn’t like them” - Your company docs agent should know you used to support Salesforce in v1, but deprecated support in v2 LLMs, surprisingly enough, do a great job understanding this. But they can only reason with the information they have, and if the vector database doesn't retrieve temporally relevant chunks, it doesn’t matter. Luckily, we solved this with Graphiti, Zep’s open-source graph library. DM if you’re interested in learning more.

I think all the hype around LLMs has caused us to overlook Small Language Models. Let me explain - you might actually realize you need them. We’re building the foundational memory later for AI agents. That means our customers expect low latency - no one wants to wait more than a few seconds for a response from an AI agent. Just imagine trying to build a sales bot that takes 15 seconds to answer your question. You’d just buy from somebody else. So the ability to do high-leverage tasks (like sales) quickly is extremely powerful. This is where smaller-language models can give you leverage. Fine-tuning for specific tasks and using accurate memory to ensure right data is provided can result in faster results and with the same accuracy as frontier models. Example models that come to mind: Microsoft Phi-3, Llama3.2 11B Just don’t feel like you have to throw the largest and most expensive model at a problem.

“What if OpenAI copies your idea?” We’re living this out right now, and let me tell you what it’s like. First, some background: LLMs have a problem with memory. You shouldn’t just send a huge data dump to GPT-4o as context - it's inefficient, and you’ll need to worry about hallucination and poor recall. We realized this in 2023, and now we’ve built the foundational memory layer for AI. Now, your agents can work intelligently with real context - extracting facts and observations from dialogue. …And OpenAI, of course, realized they needed this too. They’re developing similar features, and shipping new models. Models that will get better, faster, cheaper, with larger context windows, and better reasoning capabilities. But I’m not worried, for two key reasons. First, we built Zep to benefit from these model improvements. Second, models alone can’t solve some key enterprise use cases: - Deciding which enterprise data to feed into a prompt from a huge universe of user touch points - Building low-latency pipelines to deliver this data - Securing and governing the data with multi-tenancy, privacy compliance, RBAC, etc These are enterprise necessities that A) aren’t the domain of a model company and B) get easier as models get better. On the surface we look similar, but really, we’re playing different games. That’s why I’m not worried.

Seeing Anthropic Claude solving complex tasks using Zep AI (YC W24) for memory is fun! 💥 ➡️ Anthropic announced the Model Context Protocol yesterday, and Paul P. built an MCP server for Zep. With Zep, Claude has access to a Knowledge Graph memory spanning business data and chat history, enabling it to solve complicated tasks such as customer support tickets. 💬 + 🗄️ Developers can add business data, such as CRM tickets, emails, billing information, and more to Zep’s Knowledge Graph. With MCP, Claude can query Zep for relevant memory, assisting with queries such as “Help, I can’t log in!”. 🪄⭐️ The Zep PR is up on the MCP Server GitHub. https://lnkd.in/grkeNUBV Looking forward to other interesting projects built using MCP. If you try out the Zep Server, let us know what you think! ❤️

There's so much at stake in next week's US elections. 🇺🇸 The Russian state has tried hard to influence the outcomes, sowing disbelief and discord. The Zep AI (YC W24) team built a visual tool to explore Russia's recent activities. View the Explorer: ➡️ https://lnkd.in/g7Mh7FmC The app refers to 50+ sources, including US DOJ indictments, US and foreign government research, non-governmental organization research, and media articles. Offering users a detailed view of these malign influence operations. The Explorer uses Graphiti, Zep's open-source Knowledge Graph library. The AI Assistant was built with LangChain's LangGraph framework and FastHTML. We've spent hours delving into the data. 🔎 We hope you find it as fascinating as we did. Let us know what you think!

We missed this a few days ago! Thanks for the shout-out, LangChain!

This is a fantastic overview, Eric Vyacheslav! Thanks for the writing about Zep AI (YC W24)! ❤️ 🚀