HCP is the export format RAGMI Labs created to bridge structured human history and active AI cognition. Unlike raw data dumps, HCP ensures your personal data is immediately intelligible to any AI agent — with significance baked in from the start.
Preserves the strict hierarchy, relationships, and weights of your life data. Universally portable across platforms. Maintains the full graph of who, what, and when — machine-readable and future-proof against any single vendor.
Transforms structured data into natural language optimised for RAG ingestion. Narrative serialization with metadata injection and significance weighting — enabling immediate, high-fidelity interaction with OpenClaw and any LLM.
HCP decouples your human context from any single AI platform. Switch LLMs, switch agents, switch frameworks — your story travels with you. When future agent architectures emerge, your legacy data will still be consumable without re-authoring.
A Human RAG is an unprecedented security target. A breach can reveal psychological, relational, and emotional vulnerabilities — not just credentials. It's a honeypot problem.
RAGMI's response is a local-first architecture. All processing happens on your Mac. HCP files are stored locally — never transmitted to RAGMI servers. A physical air gap separates your identity data from any central vendor.
Data stored on RAGMI servers. Ever.
Local processing. Your Mac. Your drive.
Vendor independence. Your data outlasts any platform.
Person who owns your folio. You.
The full technical and conceptual foundations of Human RAG and the HCP standard — including ethical considerations and security analysis.