Back

AI evidence map

What is actually proven

A current-state map separates real internal/private evidence from public capability. Local, private, prototype, source-compatible, or guarded-test evidence does not automatically become public production status.

AI-assisted Project Workspace

Status: active internally for drafting, analysis, synthesis, evidence organisation, planning, and review support. Boundary: project workflow, not a public AI service and not independent decision authority.

SoulAI RC1 / RC2

Private release-candidate evidence exists. RC1 private proof passed; RC2 local guarded/full-chain verification passed. This proves a narrow bounded workflow can be engineered and exercised; it does not prove public reliability, professional quality, or production authorization.

Public SoulAI

Status: not live / not authorized for production traffic in the controlling evidence. Public expansion remains frozen until source, privacy, consent, deletion, rate-limit, reviewer, incident, and rollback gates are satisfied.

meYoo / meuus.app AI

Status: prototype / preview in the recovered AI current-reality record; not production authority. The broader meYoo idea remains a future bounded navigation/reflection interface unless a later runtime audit proves a narrower live scope.

Local model runtime

A local model runtime was verified on the workstation. This proves local inference capability, not a governed meUus Local AI product, approved retrieval, benchmark performance, privacy controls, or public integration.

AI OS / multi-agent work

Recovered code and architecture are meaningful prototype evidence. Earlier inspection found the AI OS should not be treated as production-ready. Multi-agent direction remains experimental/quarantined, and no autonomous high-stakes action is authorized.

Human and professional review

Human-review policy exists, but operations are incomplete and no continuous or guaranteed response should be promised. A qualified reviewer network or professional-grade service is not established merely because architecture describes one.

Reflection

Questions to sit with

  1. Which evidence category applies: internal, private, local, prototype, guarded test, or public production?
  2. What gate is still open before this can become public?
  3. Has a later audit changed this status?

One safe next action

Rewrite any AI claim as: Capability → Evidence → Current status → Boundary → Next gate.