Your AI should remember who you are.

The problem with cloud assistants

• Every conversation trains someone else's model
• They forget you between sessions
• They cannot hold your values, projects, or decision history
• Rate limits and price hikes break your workflow
• Your most private thoughts pass through a corporate API

The sovereign alternative

A local AI twin that runs on your own machine. It reads your memory files, counsels with your values, and keeps continuity across sessions.

What you'll build

• Personal memory layer — values, decisions, voice, and project notes in markdown
• Local inference setup — Ollama running models you choose, no API keys
• 12 calibrated prompts — calibration, morning brief, decision counsel, pattern extraction, weekly synthesis, sleep learning, and more
• Integration map — Obsidian, calendar, email, and mesh tools
• Guardian protocol — freeze-line safety for money, legal, physical, and human-welfare actions

Who this is for

• Builders who want AI continuity without cloud dependency
• Neurodivergent operators who need externalized memory and decision support
• Privacy-conscious creators, researchers, and engineers
• Anyone who has tried to make ChatGPT "remember" and found it lacking

What's inside the kit

• Sovereign AI Twin architecture blueprint
• One-hour setup playbook
• Memory templates (values, decisions, voice, projects)
• 12 prompt templates for daily use
• Integration notes for common tools
• Guardian checklist and freeze-line protocol