AI Passport — Vision
This document is the project constitution.
Mission and principles here should remain stable. Technical details live inSPEC.md,ARCHITECTURE.md, andAPI.md.
Brand
| Element | Text |
|---|---|
| Mission (primary) | One identity. Every AI. |
| Tagline (secondary) | Your AI identity. Everywhere. |
| North star | A user should be recognized by any AI, with explicit permission, without having to start over. |
Problem
Current AI systems have isolated memories. There is no portable identity that belongs to the user.
Result: repeated onboarding, repeated explanations, inconsistent AI behavior, lost productivity.
Solution
AI Passport is an open, portable identity layer. The user owns the passport. AI systems receive temporary, scoped permission to read selected parts.
Identity vs. context
| AI Passport stores | AI Passport does not store |
|---|---|
| Who you are (identity) | Full chat history |
| How you work (preferences, coding style) | Platform-specific memory blobs |
| What you are working on (active projects) | Secrets the user did not explicitly add |
| What each AI may read (permissions) | Provider-owned profile data |
Design Principles
- User ownership — The passport belongs to the user.
- Portable — Works across AI providers and IDEs.
- Permission-based — Explicit grants only; revocation is immediate.
- Local-first — Primary storage on the user’s device.
- Encrypted — Each section encrypted independently.
- Extensible — Plugins enrich the passport; consumers read it.
- Core independence — Core never depends on a specific IDE or AI company.
Long-term vision
Instead of “Create a new AI profile” on every platform, users choose Sign in with AI Passport.
The AI instantly understands who they are, how they work, and how to help — without rebuilding context every time.
See ROADMAP.md for phases. Contributor principles: MANIFESTO.md.