Format guide · .md / .txt

Markdown and plain text in the converter

Feeding Markdown to a Markdown converter sounds circular. It isn't, because most of what this tool does was never "conversion" — it's measurement and retargeting, and both apply to files that are already text.

What you get for a .md file

The element mapping

In the fileWhat happens
First # headingBecomes the document title
HeadingsSplit into addressable sections (this is what presets and anchors hang off)
Tables / figures / fences / equationsCounted for the fidelity report; content untouched
.txt (no markup)An overview section (lines, words, token estimate) + the verbatim content — honest framing for a model, no fake structure invented

Before → after

In the file

meeting notes tuesday
decided to move the deadline
carol owns the rollout
also need to fix the docs

In the Markdown

## Overview
4 lines · 17 words · ≈22 tokens — no headings detected
(QC will tell you a model sees an unstructured wall)

## Content
meeting notes tuesday
decided to move the deadline…

FAQ

Is my prose rewritten? Never — body content passes through byte-faithful; only structure is indexed and preset wrappers are added around it.

Why does my .txt score low on QC? Because it has no detectable structure — that's the report telling you a model will see a wall. Add headings (or accept the wall knowingly).

Front matter in my .md? Passes through with the body today; the Archive preset adds its own document-level frontmatter block.

Run your notes through the QC — eight checks, ten seconds.