Try it now —no sign-up needed
100% free, browser-based, your files never leave your device.
To convert a text-based PDF to editable Word, open it in ToolKnit PDF to Word, select the pages, choose original-page or A4 layout, and download the DOCX. Review the result in your editor before sharing it. The conversion runs locally. Scanned pages are not supported, and keeping the appearance of a difficult region does not make that region editable.
First check whether the PDF has usable text
Try selecting and copying a sentence from the PDF. If it produces readable words, the page may be suitable for text conversion. If the page is just a photograph or scan, this converter cannot recognize its words. A selectable but inaccurate OCR layer is not proof of a good source: copied text can be wrong even when the page looks clear.
For a mixed document, choose only the native-text pages, for example 1-3, 5. A selected page with visible content but no supported visible text stops the conversion and reports its page number. A genuinely blank page is allowed. The tool does not silently replace an unsupported scan with a picture-only Word page.
If you only need a readable picture of each page, use PDF to Image. It renders scanned pages too, without claiming to recognize their text. OCR conversion is currently unavailable in PDF to Word.
Choose original pages or an A4 document
| Choice | Useful for | What to check |
|---|---|---|
| Keep original pages | Reports, forms and documents where positions matter | Source page sizes and positions are prioritized. Text may use fixed positioning; longer edits can need manual adjustment. |
| Reorganize onto A4 | A standardized reading or editing copy | Content is fitted onto A4. Spacing and page count can change. Very wide or tall content may require original-page mode. |
Neither option is a promise of a perfectly reflowing document. PDF describes where to draw content; Word also needs paragraphs, line metrics and layout rules. Rebuilding those structures involves interpretation. A two-column page can look correct while its underlying reading order still needs review, especially before copying text, translating it or making an accessible document.
Work through the original two-page practice PDF
Download the ToolKnit practice PDF, or click Try sample PDF in the converter. This is an original synthetic example, not a customer document or an accuracy benchmark. Page 1 contains paragraphs and a small table; page 2 contains two columns and a geometric illustration.
- Load the sample and use the source preview to check both pages. The preview shows the PDF, not a rendered preview of the Word output.
- Leave the range empty to select both pages. Start with Keep original pages.
- Convert, then choose Download DOCX. Conversion prepares the file; you decide when to download it.
- Open the DOCX in your editor. Compare the headings, column order, illustration and the values 24.00, 9.00 and 33.00.
- Return to the converter and try A4 mode. Compare both versions before choosing the one that needs less editing.
What the local browser test established: both modes produced readable DOCX packages containing the expected text, numeric values and image parts. In this sample the table values remained editable, but they were represented as positioned text rather than native Word table cells. That is a useful distinction: a number that can be edited is not necessarily part of a structured table. Visual layout in your particular Word, WPS or other editor still needs checking.
What remains editable after conversion?
Paragraphs and table values
The desktop-derived engine uses text positions, drawing instructions and font information. Recognized tables can become native Word tables when their structure passes its checks. Other tables may use positioned text or image preservation. After opening the result, click inside a cell and check whether the editor offers table controls; do not infer table structure solely from its appearance.
Formulas, diagrams and uncertain regions
Complex formulas normally remain local images in this web version. Illustrations and difficult regions can also be images. The conversion report distinguishes editable characters from characters preserved in images, but those counts are not an accuracy score and do not count every graphical object. An equation visible in Word is not automatically an editable equation.
Fonts, line breaks and page geometry
Font availability and rendering differ between editors. Compare titles, superscripts, punctuation and line endings. Fixed-page output is useful for small wording changes, but expanding a short label into a paragraph may overlap nearby content. If you need major rewriting, use the conversion as a starting point and rebuild the affected section in normal Word paragraphs.
Troubleshooting without losing track of the source
| What you see | Next action |
|---|---|
| A page has no supported visible text | Exclude that scanned or image-only page, or render it with PDF to Image. |
| Invalid page range | Use source page numbers such as 1-3, 5. Endpoints must exist and ranges must run forwards. |
| Password prompt | Enter the opening password for a document you are authorized to use. It is not stored or sent to a server. Copy-restricted documents are not converted. |
| Layout or content coverage check fails | Try fewer pages to identify the problematic page. Use page images if appearance is enough; do not treat a blocked export as a successful conversion. |
| A4 does not fit safely | Switch to original-page mode and compare the result. |
| Browser stops on a complex file | Reduce selected pages or use a smaller PDF. High page complexity can hit a limit even when the file is small. |
The input budget is 256 MiB on desktop or 128 MiB on phones and low-memory devices, with up to 10,000 source pages. You can select up to 3,000 pages on desktop or 1,000 on smaller devices. The converter builds DOCX parts containing at most 1,000 or 200 source pages respectively, and reduces groups when accumulated content reaches its memory budget. These are admission budgets, not a guarantee for every PDF. Scans remain unsupported.
Privacy, downloads and the next step
The engine and its versioned dependencies download to your browser. Your PDF, its opening password and the DOCX are not sent to ToolKnit or an AI service. The page can still load fonts or send ordinary visit statistics. This workflow uses no AI key and no OCR model. A browser refresh clears the working document, so save the output before leaving.
The result is modern DOCX, not the legacy DOC format. Shared daily usage and browser safety limits apply; there is no added watermark. Once you have reviewed and edited the document, Word to PDF can help prepare a new PDF. If you only need plain text, Extract Text may be a simpler route.
Before you share the converted document
- Check names, dates, totals and other important values against the PDF.
- Review every important page in the editor that will actually be used.
- Confirm table structure, columns, images, formulas and page breaks.
- Save an edited copy separately so the original PDF remains available for comparison.
Open PDF to Word and try the practice document. Choose the result that gives you a reliable starting point, then make the final layout decisions in your Word editor.