Move documents and data without losing their meaning

Choose the result by the job it must do

Convert to an editable document when someone needs to revise the writing, to structured data when software needs the values, and to an image when the page’s appearance is the deliverable. These outputs can look similar on screen while preserving different information.

Destination task	Suitable starting path	Check before accepting the result
Edit prose in a documentation system	Document to Markdown	Heading levels, reading order, lists, links
Publish existing Markdown	Markdown to HTML	Destination rendering and allowed markup
Import rows into a spreadsheet or application	JSON and CSV conversion	Column meanings, types, identifiers, empty values
Read or troubleshoot JSON	JSON formatting	Syntax, then the recipient’s required fields
Show a fixed page in a presentation	PDF to image	Pixel dimensions, page selection, legibility

A screenshot of a table is useful for showing its appearance, but it is not the same deliverable as its rows and values. Similarly, Markdown text may remain editable while losing the original page breaks and precise positioning. Agree on the destination task before producing several unnecessary versions.

Create a small review sample from the source

Keep the original file. Pick sections or records that are likely to expose a conversion problem, rather than reviewing only the opening paragraph or the first easy row.

For a document, choose a heading, a numbered list, a table, a footnote, and a page with multiple columns if one exists. For data, choose a leading-zero identifier, a genuine zero, an empty value, a non-English name, and a field containing a comma or a line break.

The following is a review plan for a hypothetical report, not a claim that a particular uploaded report was tested:

Source item	What must remain true
Section titled “Delivery exceptions”	Its paragraphs remain beneath that heading
Account identifier 0012	It remains the same identifier in the destination
Change of -0.5 percentage points	The minus sign and unit remain attached to the value
Table note applying to one row	The note still clearly identifies that row
Numbered procedure	The sequence and nesting remain understandable

This makes a successful review concrete. A file opening without an error is only the beginning of the check.

Inspect document structure before polishing its appearance

Tooliv’s DOCX conversion uses Mammoth to produce HTML before converting that result to Markdown. Mammoth’s documentation describes an emphasis on document meaning rather than exact visual styling. Complex documents can therefore need manual cleanup.

A source paragraph styled as a heading gives the converter a clearer structural signal than text that merely looks large and bold. After conversion, copy the result into your Markdown editor and check the hierarchy. A simple section might be written as:

## Delivery exceptions

### Missing information

1. Check the account identifier.
2. Confirm the delivery address.

This is a hand-written structural example, not output from an executed DOCX conversion. Match the heading depth to the surrounding document rather than applying this exact hierarchy everywhere.

Tables need their own review. GitHub Flavored Markdown’s table syntax has a single header row and pipe-separated cells. A complex source table with merged headings, nested content, or explanatory blocks may need to be simplified into several tables or a table plus prose. Do not assume that a readable preview proves every cell is still associated with the right heading.

For PDFs, Tooliv’s document converter extracts an existing text layer. It does not perform OCR. A scan containing only images needs a separate text-recognition step; a scan with an existing text layer still needs its text and reading order checked. If you only need a page illustration, the PDF image guide explains that separate workflow.

Review table meaning separately from table shape

JSON to CSV accepts pasted text or one UTF-8 JSON, CSV, or TSV file, up to 5,000,000 bytes. Select the comma, semicolon, or tab delimiter and inspect the table preview, which shows at most ten data rows and twelve columns. This sample does not replace a check of the full output. A valid conversion can preserve rows while changing types or collapsing empty-value distinctions; the JSON-versus-CSV walkthrough shows an actual conversion-function example.

For a spreadsheet handoff, record what one row represents, whether an amount is a total or a unit price, and how missing values should be handled. A customer identifier should not become a quantity merely because it contains digits.

Commas, quotes, and line breaks inside fields also affect how a CSV is read. RFC 4180 documents the common quoting conventions. If columns split unexpectedly or names display incorrectly, follow the CSV import troubleshooting guide before changing the source values.

The JSON formatter helps inspect syntax and nesting. It does not know whether a required business field is missing or whether a date represents the correct time zone. Keep important identifiers as strings in the source, and validate the result against the receiving application’s rules.

Finish in the actual destination

Open the downloaded or copied result in the editor, spreadsheet, or publishing system that will use it. Compare the review sample again. Different import rules or Markdown support can change the result after the converter has finished.

Record the source filename, selected output, and any manual changes. If a check fails, return to the source or documented conversion step instead of repeatedly editing unexplained differences into the final file. Once the sample is sound, check the complete document or dataset for missing sections, unexpected row counts, and incomplete output before handing it over.