What If Your AI Agent Could Remember a Lie?

Your AI Has a Memory. Who Protects It From Lies? 🛡️

Imagine an AI agent that remembers everything you tell it—but can't tell the truth from a carefully planted lie.

That's the problem that inspired Palisade.

Built for Graph Hacks by WeMakeDevs × FalkorDB, Palisade explores a simple idea: AI agents shouldn't just remember. They should know where their memories came from.

🧠 Why FalkorDB?

A memory without context is just text. Palisade uses FalkorDB to connect memories to their sources and trace how one memory leads to another. Combined with graph-enhanced RAG and a multi-agent security workflow, it helps surface suspicious instructions and produce evidence-backed answers.

⚔️ The bigger idea

What if an AI's memory worked less like a notebook and more like an immune system—examining suspicious information instead of blindly trusting it?

That's what I'm exploring with Palisade.

Built with FalkorDB, FastAPI, React, Sentence Transformers, and local Qwen 2.5 3B.

🔗 Explore the project: https://github.com/OMKHEDIKAR23/Palisade

Because the future of AI isn't just about what agents know. It's about what they have reason to trust.