Path-only vs full enhance

UBLX separates fast indexing from deep metadata. Understanding the split keeps large trees usable.

Defaults

Mode	On snapshot	Catalog contains
Path-only (default)	Nefaxer walk only	Paths, categories, filetype hints
Full enhance	Nefaxer + ZahirScan batch	Zahir JSON for previews, Templates, Metadata, Writing

enable_enhance_all = false is the default. Turn on full enhance globally, per subtree with [[enhance_policy]], or file-by-file in the TUI.

When to stay path-only

• First pass on a large directory — see filetype mix without waiting on every file
• Trees where you only need Delta / duplicate / path navigation
• CI that only needs a snapshot or export skeleton before selective enhance

When to batch enhance

• Smaller projects where you want rich previews immediately
• Subtrees you always inspect (e.g. src/) — use policy = "auto" on that prefix
• Headless export pipelines — prefer --full-snapshot or tuned policies before --export

On-demand enhance

For policy = "manual" prefixes (or path-only rows), use Enhance with ZahirScan from quick actions or multi-select. Good when only a few files need depth.

Tradeoffs

Approach	Speed	Depth	Best for
Path-only	Fastest	Shallow	Exploration, huge trees
Policy auto on prefix	Medium	Deep in subtree	Monorepos, mixed content
enable_enhance_all	Slowest	Deep everywhere	Small roots, full export
On-demand	Pay per file	Deep where needed	Archives, occasional deep dives

Next: Enhance policies, Headless snapshot + export.