copy-tag-link
Shared AI Memory
Welcome to Copy-Tag-Link, or CTL for short. A file-first ingestion standard for shared AI memory.
When you build AI knowledge systems, you start with rich documents containing maps, columns, and data tables.
Vectors Chunk
Traditional RAG systems shred documents like scrap paper. They break down layout structures and generate raw float coordinate matrices.
The core structure is lost, image contexts dissolve, and exact coordinate/table data are replaced by approximations.
Markdown Flattens
Converting rich layouts directly to Markdown compresses multi-column layouts into a vertical stream, scrambling chronological order.
Table cells lose boundaries, and spatial components like maps are reduced to flat, detached text captions.
Copy-Tag-Link Preserves
We preserve the original source as evidence, copy out layout assets like images and tables, tag them with provenance, and link them to semantic HTML views.
The output remains human-readable, machine-validatable, and fully independent of any one model.
OKF Indexing
These parsed components are cataloged using OKF cards: markdown files with standard YAML metadata tags.
The OKF cards act as portable index entries, making your rich HTML segments and assets instantly searchable and accessible to any database or AI agent you choose.
