
IIIF built five rings. Each solves access. None solves permanence.
We built cathedrals of knowledge on rented land.
museum.org/image.jpg
Server dies → content gone
Fragile. Centralized. Ephemeral.
bafybeigdyrzt5sfp7udm7hu76uh7y26nf3efuylqabf3oclgtqy55fbzdi
The content is its own address
Resilient. Distributed. Permanent.
No one owns it. Everyone can serve it.
Files distributed across network nodes — any node with a copy can serve it
Hash derived from content itself: bafy...7xk2
Content is its own address — change one pixel, hash changes
If I create an annotation and store it on IPFS, anyone with that CID can retrieve it — not from my server, but from any node on the network that has a copy
Annotation → CID generated
Requests same CID instantly
Pins content as steward
Capture full state at moment of engagement
Generate cryptographic identifier (CID)
Store across peer network
Translate to/from 20 canonical elements
Access from any participating node
Scholarly intent preserved
Institutional metadata
Authorship & credit
Scientific & archival
Derived from established standards — enough to round-trip between formats without data loss
Nested metadata, multiple formats. Translation preserves intention, not just information
Not an extension—a substrate beneath. Codex sits as a foundational layer underneath IIIF, Dublin Core, MODS, VRA Core, and Four Corners.
Full round-trip compatibility with all major metadata standards
The hinge connects without coupling—preserving institutional autonomy
A common foundation that makes every platform survivable
Pin your own content = abandonment with extra steps
Stores Yale's snapshots
Stores small museum's content
Stores Harvard's metadata
Storage is the currency. Participation is the price of preservation.
The substrate that makes all platforms survivable.
Curiosity is Collaboration
email sonia@thetechmargin.com
IG: @TheTechMargin
LinkedIn: Connect on LinkedIn
Codex Protocol