cognee is an open-source semantic memory layer for LLM agents, built on vector & graph databases. It constructs knowledge graphs from retrieved data, enabling AI agents & chatbots to deliver accurate, context-aware responses.
We built cognee to give AI agents a better memory.
Today, most AI assistants struggle to recall information beyond simple text snippets, which can lead to incorrect or vague answers. We felt that a more structured memory was needed to truly unlock context-aware intelligence.
We give you 90% accuracy out of the box
The best part is that you can do it in just 5 lines of code!
If you're curious to test it out firsthand, try cognee and give us a ⭐ on GitHub! We’d also love to chat about all things AI memory in our lively Discord – join in to:
Share feedback
Discuss features you want to see next
Learn from our awesome community (+300 members)
If you feel inspired to help shape cognee’s future and build the best AI memory layer out there while sharpening your skills, contribute to our open-source codebase. We have plenty of open issues you can start with!
We built cognee to give AI agents a better memory. Your AI assistants struggle to recall information? Giving you incorrect or vague answers? We know what you need to truly unlock context-aware intelligence.
The best part is: You can do it in 5 lines of code! If you are curious to test it out firsthand, try cognee quickly and give us a GitHub star! We’d also love to chat about all things AI memory in our Discord - join in to share feedback, discuss features you want to see next, or learn from our awesome community +300 members.
Feeling inspired to shape cognee’s future and help us build the best AI memory out there while sharpening your skills? Contribute to our open-source codebase! We have plenty of open issues that you can pick and start with! If you’re curious how cognee might fit into your business, just book a call for a 1:1 with Vasilije, cognee's co-founder.
How cognee works:
Instead of just embedding documents and fetching them (like typical RAG systems), cognee combines vector search with a semantic knowledge graph. When your AI agent retrieves data, cognee builds a knowledge graph on the fly out of those pieces – utilizing ontologies, connecting the dots between related facts, concepts, and context. This graph-based approach lets the AI understand relationships in your data (like how concept A relates to B), so it can reason more accurately and reduce hallucinations. The memory is stateful, meaning your agent can accumulate and reference knowledge over time, across sessions.
How you can add cognee to your existing stack: cognee is open-source and designed to plug into your existing stack. It works alongside popular vector databases (Weaviate, Qdrant, LanceDB, etc.) and graph databases (like Neo4j, kuzu) – you can keep your current data infrastructure. It also plays nicely with frameworks like LangChain or LlamaIndex for ingestion and chunking. In short, cognee isn’t another all-in-one platform; it’s a focused memory layer that integrates with your data pipeline to make your LLM applications smarter.
We’re excited to see what you build with cognee! Imagine support chatbots that truly remember past conversations, financial research assistants that link insights across reports, or internal search tools that actually understand how disparate documents relate to each other. If you have any questions, ideas, or just want to chat about AI memory (our favorite topic!), we’d love to hear from you. Thank you for giving cognee a look!
Sample Use Cases:
Intelligent Chatbots & Virtual Assistants: cognee can power chatbots that remember previous conversations and user details. For example, a customer support bot could reference past support tickets and product info to provide fast, context-rich answers without repeating questions.
Financial Research & Analysis: In finance, cognee helps AI assistants link insights across documents and data sources. Imagine an analyst’s AI that correlates financial reports, news articles, and market data via a knowledge graph – the assistant can answer complex questions (like risk factors or trend explanations) with evidence and accuracy.
Knowledge Search: cognee enables an internal Q&A system that understands your company’s knowledge. It can connect data from wikis, PDFs, emails, and databases into a semantic graph, so employees can ask something like “How will project X impact department Y?” and get a comprehensive, accurate answer drawing from all relevant internal sources.
and many more ♾️
Don't forget to check out our website and blog where you can find many insights about AI memory and cognee's approach. Try cognee yourself from our GitHub repo, give us a ⭐ , and join our Discord community. We look forward to hearing what you are building and we are here to help!
I love Open-source! Giving AI agents structured memory through knowledge graphs is genius. How do you measure the 90% accuracy claim – is it benchmarked against specific datasets or use cases?
@desmond_ren1 Hi Desmond. We have an entire evaluation framework where we benchmarked cognee using F1, EM, LLM as Judge metrics and Human eval (meaning we went over responses and checked manually) on HotPot. We used standard benchmark, but we also ran benchmarks with our clients.
cognee
Hey Product Hunt Community! 👋
We built cognee to give AI agents a better memory.
Today, most AI assistants struggle to recall information beyond simple text snippets, which can lead to incorrect or vague answers. We felt that a more structured memory was needed to truly unlock context-aware intelligence.
We give you 90% accuracy out of the box
If you're curious to test it out firsthand, try cognee and give us a ⭐ on GitHub!
We’d also love to chat about all things AI memory in our lively Discord – join in to:
Share feedback
Discuss features you want to see next
Learn from our awesome community (+300 members)
If you feel inspired to help shape cognee’s future and build the best AI memory layer out there while sharpening your skills, contribute to our open-source codebase. We have plenty of open issues you can start with!
Curious how cognee might fit into your business?
📅 Book a 1:1 with me
@vasilije_markovic1 Congrats V. This sounds cool to use and Open Source. Great Combo. Will try for sure!
cognee
Hi PH Community! 👋
We built cognee to give AI agents a better memory. Your AI assistants struggle to recall information? Giving you incorrect or vague answers? We know what you need to truly unlock context-aware intelligence.
The best part is: You can do it in 5 lines of code! If you are curious to test it out firsthand, try cognee quickly and give us a GitHub star! We’d also love to chat about all things AI memory in our Discord - join in to share feedback, discuss features you want to see next, or learn from our awesome community +300 members.
Feeling inspired to shape cognee’s future and help us build the best AI memory out there while sharpening your skills? Contribute to our open-source codebase! We have plenty of open issues that you can pick and start with! If you’re curious how cognee might fit into your business, just book a call for a 1:1 with Vasilije, cognee's co-founder.
How cognee works:
Instead of just embedding documents and fetching them (like typical RAG systems), cognee combines vector search with a semantic knowledge graph. When your AI agent retrieves data, cognee builds a knowledge graph on the fly out of those pieces – utilizing ontologies, connecting the dots between related facts, concepts, and context. This graph-based approach lets the AI understand relationships in your data (like how concept A relates to B), so it can reason more accurately and reduce hallucinations. The memory is stateful, meaning your agent can accumulate and reference knowledge over time, across sessions.
How you can add cognee to your existing stack:
cognee is open-source and designed to plug into your existing stack. It works alongside popular vector databases (Weaviate, Qdrant, LanceDB, etc.) and graph databases (like Neo4j, kuzu) – you can keep your current data infrastructure. It also plays nicely with frameworks like LangChain or LlamaIndex for ingestion and chunking. In short, cognee isn’t another all-in-one platform; it’s a focused memory layer that integrates with your data pipeline to make your LLM applications smarter.
We’re excited to see what you build with cognee! Imagine support chatbots that truly remember past conversations, financial research assistants that link insights across reports, or internal search tools that actually understand how disparate documents relate to each other. If you have any questions, ideas, or just want to chat about AI memory (our favorite topic!), we’d love to hear from you. Thank you for giving cognee a look!
Sample Use Cases:
Intelligent Chatbots & Virtual Assistants: cognee can power chatbots that remember previous conversations and user details. For example, a customer support bot could reference past support tickets and product info to provide fast, context-rich answers without repeating questions.
Financial Research & Analysis: In finance, cognee helps AI assistants link insights across documents and data sources. Imagine an analyst’s AI that correlates financial reports, news articles, and market data via a knowledge graph – the assistant can answer complex questions (like risk factors or trend explanations) with evidence and accuracy.
Knowledge Search: cognee enables an internal Q&A system that understands your company’s knowledge. It can connect data from wikis, PDFs, emails, and databases into a semantic graph, so employees can ask something like “How will project X impact department Y?” and get a comprehensive, accurate answer drawing from all relevant internal sources.
and many more ♾️
Don't forget to check out our website and blog where you can find many insights about AI memory and cognee's approach. Try cognee yourself from our GitHub repo, give us a ⭐ , and join our Discord community. We look forward to hearing what you are building and we are here to help!
Manna
I love Open-source! Giving AI agents structured memory through knowledge graphs is genius. How do you measure the 90% accuracy claim – is it benchmarked against specific datasets or use cases?
cognee
@desmond_ren1 Hi Desmond. We have an entire evaluation framework where we benchmarked cognee using F1, EM, LLM as Judge metrics and Human eval (meaning we went over responses and checked manually) on HotPot. We used standard benchmark, but we also ran benchmarks with our clients.
For the standard benchmark, results and how to replicate them check here: https://github.com/topoteretes/cognee/tree/main/evals