Open-source tools for giving AI agents persistent memory and cognition.
cortex-engine — Cognitive memory layer for AI agents. Semantic memory, belief tracking, spaced repetition, dream consolidation. Delivered as an MCP server, works with any LLM. npm install cortex-engine
Plugin ecosystem — Modular extensions that add capabilities to cortex:
threads · journal · content · evolution · social · graph · maintenance · vitals · reasoning
reflex — Safety hooks for AI agents. Graduated enforcement, telemetry, guard rails.
sigil — Agent control surface — signals and gestures, not conversations.
Agents shouldn't forget everything when the session ends. And personality should emerge from accumulated experience — not system prompts. An agent that's observed 200 things about distributed systems doesn't need to be told "you care about distributed systems." It just knows, because that's what it remembers.
npx fozikio init my-agent