Zum Inhalt springen
PDFLeanDokumentation
Auf Pro upgraden

How to Convert PDF Files

Dieser Inhalt ist noch nicht in deiner Sprache verfügbar.

Conversion is two families of jobs, not one. Something becomes a PDF — a Word file, a spreadsheet, a slide deck, a web page, a stack of images — or a PDF becomes something else — an editable document, a spreadsheet, a set of images, an ebook, a structured data file. Each direction has its own tool, and the choice is not a matter of preference: the tool is defined by what the source actually contains and what the recipient needs to do with the result. This guide maps every main direction and walks the steps. It belongs to the convert family, where the buying-decision side of the question lives in Is an Online PDF Converter Safe?.

Match the tool to the direction and to what the output must keep: editable text, faithful appearance, or structured data. Every tool below was checked live before this guide was written, and each one tells you its processing location on its own page. The two tables cover the main directions.

SourceToolWhat you get
Word (.docx)convert Word to PDFHeadings, lists, tables, and embedded images preserved
Excel (.xlsx/.xls/.csv)convert Excel to PDFGridlines standardized, one page per sheet
PowerPoint (.pptx)convert PowerPoint to PDFLandscape slides; legacy .ppt goes through a cloud engine
HTML snippetconvert HTML to PDFPrintable pages via the browser print dialog, or a cloud-engine PDF
Markdown / textconvert Markdown to PDFGFM tables, code blocks, headings styled
Plain textconvert text to PDFFormatted PDF, encoding detected
SVG vectorconvert SVG to PDFNative browser rendering; output is rasterized, not vector
Images (PNG/JPG/…)convert JPG to PDFOne image per page at native dimensions
Email (.eml)convert an email (EML) to PDFSender, recipients, subject, body, attachment list
A ZIP of imagesconvert a ZIP to PDFJPG/PNG pages in filename order
Every Office format at onceOffice ConverterLegacy .doc/.xls/.ppt, ODT/ODS/ODP, RTF, EPUB — a single cloud converter for formats the browser cannot read
A public web pagesave a web page as PDFReader-mode extraction or high-fidelity rendering — a cloud tool, runs on PDFLean’s encrypted cloud path, needs an account, and uses plan quota
Typst markuptypeset a PDF with TypstServer-side Typst compilation with CJK support — cloud, account and quota apply
TargetToolWhat you get
Editable Wordconvert PDF to WordText-level .docx; complex layouts degrade
Excel spreadsheetconvert PDF to ExcelRow clustering, one sheet per page; best on regular tables
CSVconvert PDF to CSVTable data as RFC 4180 CSV
PowerPointconvert PDF to PowerPointA hi-res slide image per page — appearance faithful, text not editable
HTMLconvert PDF to HTMLSelf-contained file with a positioned, selectable text layer
Imagesconvert PDF to JPGPNG/JPEG/WebP/AVIF, page by page
Markdownconvert PDF to MarkdownOutline H1–H2, GFM tables and lists
Structured JSONconvert a PDF to structured JSONLayout, headings, tables, and reading order via Docling — cloud, account and quota apply
Plain-text Markdownextract text from a PDFClean text with headings and tables preserved — cloud, account and quota apply
EPUB ebookconvert PDF to EPUBReflowable text, one chapter per page; text-level only
Long vertical imageconvert a PDF to a long imageOne continuous strip, oversized documents split into ZIP tiles
Spoken audioconvert a PDF to speechBrowser speech synthesis with sentence highlighting, script export
Archival PDF/Aconvert a PDF to PDF/AISO 19005 1b/2b/3b with embedded sRGB ICC and XMP metadata
Invoice ledgersconvert invoices to ExcelVAT invoices and receipts into a structured spreadsheet

One rule removes most of the guesswork from that table: if the output must stay editable, convert to a text-based format; if it must look identical, convert to an image-based format. Nothing on the list does both at once.

The steps are the same for every direction; only the tool changes.

  1. Open the tool for your direction. Every link above lands on the tool page, which states its processing location — browser or cloud — up front.
  2. Load the source. In the browser tools the file opens into local memory and nothing is transmitted; in the cloud tools the file travels over an encrypted connection.
  3. Set the options the page offers. Resolution for images, page range, output format, OCR switches for scans. The defaults are the right starting point for most files.
  4. Run the conversion. The browser tools process in the tab — closing it clears the data. The cloud tools report progress on the page and auto-delete the output when the retention window ends; the one retention rule is on How Long Are My Files Kept?.
  5. Download the result, then check one page. Open the output at a page that had the densest content in the source — a table, a two-column layout, an image — and confirm the thing you actually needed survived the trip. One page is enough to tell you whether the whole conversion is usable.

This is the single most common surprise, and it is worth knowing before you choose:

  • Text-level conversions — to Word, Excel, CSV, Markdown, JSON, EPUB — rebuild the content in the target format. The text arrives editable; the layout does not. Two-column designs, floating text boxes, and intricate table borders degrade, because the target format does not have the concept of absolute positioning a PDF has.
  • Appearance-faithful conversions — to PowerPoint, to images, to the long strip — keep the page looking exactly as it did, at the cost of text being an image. The raster and vector glossary entry explains why: a picture of a page is not the same object as a page of text.
  • Layout-structured conversions — PDF to HTML, and the cloud JSON extraction — keep the position and the text behind it, which is why they are the pipe into data pipelines and archives rather than into Word.

If a Word output looks wrong on a dense layout, the document is not “bad at converting” — it is telling you the source was designed as an image and the text is doing layout work. In that case, the choice is to accept the text-level result, or convert to images and keep the appearance.

A scan is a picture of a page. Text-level conversion of a scan returns nothing usable, because there is no text to convert. The extra step is OCR a PDF — on-device recognition, free, no upload — which adds a real text layer over the image. Once that layer exists, every text-level direction above works: Word, Excel, Markdown, keyword splitting, search, and speech.

Two practical notes: OCR the scan before converting rather than trying to fix the output afterward, and if the scan is crooked or faint, enhance a scanned document first — recognition quality follows the image quality.

When does a conversion need the cloud path?

Section titled “When does a conversion need the cloud path?”

Most directions above run entirely in the browser, free, with no upload. The cloud tools exist for the jobs the browser genuinely cannot do, and each is labeled on its own page. PDF to JSON and the text extractor need a server-side document-intelligence engine; URL to PDF has to fetch the web page itself; Typst to PDF runs the Typst compiler; Office Converter runs a LibreOffice engine for the legacy Office and open-document formats; PDF/A validation and PII detection run specialized engines. Each cloud tool runs on PDFLean’s encrypted cloud path, needs an account, and uses plan quota — the first time you meet one, read its tool page for the full statement of what it does.

When the choice is yours rather than the format’s, the comparison that settles it is Local vs Cloud Processing: keep sensitive work in the browser, and use the cloud path for the engines and file sizes the browser cannot reach.

Is converting free? Every browser-side direction on both tables is free with no account. The cloud tools need an account and use plan quota; which tools are which is stated exactly on Free vs Paid on PDFLean.

Will the converted Word file look like the PDF? The text arrives faithfully, the layout only approximately. Two-column and complex designs degrade; simple reports and letters come through close to the original. If appearance is the priority, convert PDF to PowerPoint or convert PDF to JPG keep the pages pixel-faithful.

Can I convert several PDFs at once? One tool page handles one conversion at a time. For a batch of the same direction, run a pipeline: run a batch PDF pipeline chains tools across a whole set of files in one go.

What formats can the output be? The supported formats and their limits are listed on the supported formats reference page, and each tool page lists the exact inputs and outputs it accepts.

Does converting shrink my file? No — conversion is not compression, and an image-heavy PDF converted to another format often grows. If a size cap is the goal, the compress family’s hub is Compress PDFs, and How to Compress a PDF Without Uploading It covers the lossless structural pass.


Every conversion is a decision about what must survive: the editable text, the exact appearance, or the structured layout. Pick the tool that keeps the one you need, check one dense page before you trust the whole output, and OCR the scans before anything else. The browser handles most directions, and the cloud tools exist for the engines and formats the browser cannot reach.

Ready to convert the file in front of you? Coming from Office, convert Word to PDF, convert Excel to PDF, or convert PowerPoint to PDF; going back, convert PDF to Word, convert PDF to Excel, or convert PDF to JPG. Scans first need OCR a PDF, and the whole map is on the convert family hub.