Map Extract
Free tool

Parse text out of almost any document

One parser for the files a small business actually receives: Word and PDF documents, Excel and CSV sheets, PowerPoint decks, saved emails and mailboxes, EPUB e-books, and HTML or XML pages. Choose several at once, switch between views, and export the text or its structure.

Runs in your browser. Nothing you enter is uploaded.

What it reads

FamilyFormatsExtra views
DocumentsPDF, DOCX, DOC, ODT, RTF, TXT, MDMarkdown, outline of headings, tables
SpreadsheetsXLSX, XLSM, XLSB, XLS, ODS, CSV, TSVCells as CSV, formulas, comments
PresentationsPPTX, PPT, ODP and their show and template variantsOutline, speaker notes
EmailEML, EMLX, MSG, MBOXHeaders and body, email addresses, attachment list
E-booksEPUB, FB2Chapter by chapter
Web and XMLHTML, XHTML, MHTML, XML, SVG, RSS and moreLinks, XML paths

Parsing files

  1. Pick a starting view

    The Extract buttons preset Plain text, Markdown, Outline or JSON. You can change view after a file is read.

  2. Choose the files

    Select one file or many. Each is read in turn and appears as a chip; click a chip to show that file, or remove it.

  3. Check the figures

    For the selected file you see its word and character counts, what it contains (pages, sheets, slides, messages or chapters) and the format that was detected.

  4. Switch views

    The View control lists only the views that make sense for that file, so formulas appear for a workbook and speaker notes for a deck.

  5. Export

    Download the current view as TXT, MD, JSON or CSV, copy it, or export the file's tables, slides, messages or paragraphs to Excel or CSV. With several files, All as ZIP saves every one in the same view.

A supplier's terms in Word

This test file is a one-page Word document with a title, two subheadings, two short paragraphs each and a small price table. The parser reports 54 words and offers Plain text, Markdown, Outline, Tables and JSON.

The Markdown view keeps the heading levels and turns the Word table into a Markdown table, which is the most useful form for pasting into notes or giving an AI assistant context. The Outline view lists just the headings.

Input
terms-v1.docx (Word)
Heading 1: Wholesale terms 2027
Heading 2: Minimum order, Payment
Table: Range / Trade price
Result
# Wholesale terms 2027

Kiln & Co Pottery supplies stockists in the South West on the terms below.

## Minimum order

The minimum first order is 24 pieces. Repeat orders have no minimum.

## Payment

Invoices are payable within 30 days. Contact accounts@kilnandco.example with questions.

| Range | Trade price |
| --- | --- |
| Stoneware mug | £9.00 |
| Pouring jug | £16.00 |

Limits worth knowing

A PDF that is a scan has pictures of words, not text. When a PDF yields almost no text per page, the parser says so and suggests running it through OCR first. PDFs with columns or unusual layouts are read in the order the text sits on the page, which is usually right but can interleave columns.

Some formats are not read at all, and the page tells you why: Outlook PST and OST data files, Apple Pages files and compiled CHM help files. Export those to MBOX or MSG, DOCX or PDF, or HTML first. A file whose extension does not match its contents is read as what it really is, with a note saying so.

Questions people ask

What is a document parser?

A document parser opens a file and pulls out its content in a structured form: the text, but also headings, tables, sheets, slides or messages depending on the format. This one runs in your browser, reads more than twenty formats, and lets you export the result as text, Markdown, JSON, CSV or an Excel workbook.

Can it read scanned PDFs?

Not directly. A scanned page is an image, so there is no text inside the file for a parser to read. When a PDF has very little text per page, a note says it looks scanned. Run it through an OCR tool first, then parse the result or copy the recognised text.

How many files can I parse at once?

As many as you choose. They are read one after another and listed as chips you can switch between. When more than one has been read, All as ZIP saves every file's text in the current view, which is handy for a folder of supplier documents or a batch of saved emails.

Why does it say a file belongs to another parser?

That message comes from the single-format parsers on this site, such as the spreadsheet or presentation extractors, when they are given a file of another family. This all-formats parser accepts every family, so if you see that message, use this page or the parser for that format.

Where your data goes

Files are opened and parsed by your browser, with libraries loaded only when a format needs them. No document, and no text taken from one, is uploaded.

Checked against the tool on

Top