A Minimal Verification Layer for Public Documents

I’ve been thinking about a simple question: What if verification infrastructure should be separate from content itself? Most systems host, mirror, or interpret information. That mixes storage, meaning, and validation into one layer. I’m experimenting with a minimal, non-custodial model: Records SHA-256 over official bytes Stores URL + fingerprint + deterministic result Does not host or mirror documents Same input → same output Verification failures are preserved, not deleted It’s not a truth engine. It’s a byte-level sameness recorder. I’m curious whether this kind of separation makes sense as infrastructure. Accuracy (4) = Truth (2) = Life (4)