Readable Markdown for documents, PDFs, spreadsheets, HTML, and knowledge-base workflows.
support@mdforall.com
Markdown For All
Conversion limits
A centralized, versioned explanation of what the public converter can preserve, what it cannot reproduce, and when another workflow is safer.
What Markdown can preserve
Markdown is strong at durable reading structure: headings, paragraphs, lists, links, block quotes, fenced code, and regular tables. It is designed for editing and portability, not exact reproduction of coordinates, fonts, page breaks, animations, workbook logic, or interactive web behavior.
Format-by-format limits
| Format | Typical strengths | Known limits |
|---|---|---|
| Paragraphs and some inferred headings | Unicode mapping, punctuation, non-Latin characters; Nested indentation and visual grouping; Merged cells, form layouts, totals, column order | |
| Word documents | Usually become Markdown headings | Manually bolded pseudo-headings may remain paragraphs; Deep nesting and custom numbering can flatten; Merged cells, nested tables, and complex widths |
| Excel spreadsheets | Regular cell values | Formatting-dependent meaning and comments; Multi-row headers, merged cells, hidden rows/columns; Formulas, dependencies, macros, recalculation |
| PowerPoint slides | Section headings | Untitled slides and decorative text; Position-based grouping and deep custom indentation; Timing, presenter cues, hidden review comments |
| HTML | h1–h6 and readable text | CSS-only visual hierarchy; JavaScript navigation and inaccessible click handlers; Interactive widgets and syntax behavior |
| CSV to a Markdown table | One descriptive header row | Missing, duplicated, or blank column names; Embedded newlines, inconsistent quoting, delimiter collisions; Legacy encodings and replacement characters |
| JSON | Labels, values, booleans, nulls | Data types are presented as text; Mixed item shapes and order assumptions; Deep hierarchy can become verbose or ambiguous |
| XML | Section labels and record fields | Generic tag names may need human-friendly explanations; Namespaces, IDs, and inherited meaning; Optional fields and heterogeneous child elements |
| plain text | Paragraph boundaries | Accidental line wrapping and hard returns; Inconsistent symbols and nesting; Tabs, fixed-width columns, copied page layout |
Sensitive-file guidance
The public workflow is intended for public, redacted, already-shareable, or otherwise approved low-risk material. Contracts, customer exports, health information, bank statements, HR records, credentials, private source code, legal files, and regulated documents should use an approved private or self-hosted workflow.
Manual review checklist
- • Compare headings and paragraph order with the source.
- • Verify every name, date, amount, unit, identifier, clause, and total.
- • Check list nesting, table headers, row alignment, links, images, code fences, and non-Latin characters.
- • Remove repeated headers, footers, page numbers, navigation, banners, and unrelated boilerplate.
- • Keep the original file as evidence when exact layout, signatures, formulas, charts, or machine semantics matter.
Version
Last reviewed July 16, 2026. Applies to public release 2026-07-16-geo-r7. See the changelog for material updates.