Cortex
Knowledge.
Memory.
Intelligence.
A graph engine that indexes your vault, learns from every interaction, and gives your AI persistent memory.
Cortex turns your knowledge base into a graph that AI can walk, search, and remember. No cloud. No embeddings. No API calls. A single binary that runs on your machine and gives your AI the tools to understand your knowledge — not just retrieve it.
Three Layers
Graph
Indexes every note, link, embed, and tag. Walks the link graph with pruning heuristics. Scores hubs by what the AI actually uses.
Memory
Neural synapses learn which notes matter together. Semantic memories persist across sessions. A user profile builds itself from how you work.
Intelligence
Discovers hidden connections. Suggests missing links. Validates note health. Understands your topic taxonomy. Gets smarter the more you use it.
How It Works
1. Index
Point it at your knowledge base. Cortex parses every file — tags, links, embeds, relationships. Builds a full link graph in memory.
2. Walk
Your AI uses graph tools to explore. Smart traversal follows the strongest connections, prunes noise, and delivers exactly the context that matters.
3. Learn
Every walk strengthens connections between co-accessed notes. Every recall makes a memory stronger. Every session teaches the persona what matters to you.
4. Remember
Facts, decisions, insights stored as files you can read and edit. Memories decay naturally over time — what you don't use fades. What you revisit gets stronger.
Zero Dependencies
No cloud. No database. No API keys. No embeddings. No runtime.
A single binary under 5MB. Your data never leaves your machine.
Want Cortex for Your Team?
Cortex ships free with the Araptus Client Ecosystem. Standalone licenses available for teams that want the graph engine without the full stack.
Get in Touch