Extract text-based PDF pages into downloadable Markdown with page selection, visible lists, optional headings and simple table mode.
LOCAL FILE CONVERSION
Open local files, review the result and save it privately in your browser.
PDF ≤25 MiB / 200 source pages · Extract ≤30 pages · 200,000 UTF-8 output bytes · Drag and drop · Browser only; no uploadUpload PDF → Extract Markdown → Review → Download. Text, font information and page geometry guide structure detection. Everything stays in your browser.
Re-extraction replaces source edits. You can also correct uncertain tables directly in the editable Markdown below.
Enter your values locally; no inputs are uploaded.
Check validation feedback before reviewing the browser-generated result.
Review the explanation, copy supported outputs or reset the workspace.
Choose a supported local file, review compatibility and output settings, inspect actual output previews or text, then download.
Extract text-based PDF pages into downloadable Markdown with page selection, visible lists, optional headings and simple table mode. Best for horizontally oriented single-column text. Rows follow visual top-to-bottom/left-to-right positions; gaps suggest paragraph breaks, and visible list markers become Markdown syntax. Default output does not invent headings. Opt-in heading inference uses conservative size cues and leaves uncertain short pages unstructured. Table mode reuses simple aligned-table extraction and treats each selected page’s first row as a header; review every cell. Complex layouts can require cleanup. Text syntax is escaped; images are not exported. No OCR/AI. Limits: 25 MiB PDF, 200 source pages, 30 selected pages and 200,000 UTF-8 output bytes.
Source files, filenames and document content stay in browser memory. No conversion uploads, external processing services, file persistence or content analytics are used. Cancel/reset/unmount discards pending work and owned URLs; ordinary page requests may still be logged.
Text-based, horizontally oriented, single-column pages work best. Scans without text need OCR elsewhere. Complex columns and layouts can produce reading-order ambiguities and require manual cleanup.
Visible list markers become Markdown list syntax. Heading inference is opt-in and relies on clear font-size differences; uncertain short pages stay plain text. Default extraction avoids inventing headings.
Optional table mode reuses the aligned text-table extractor and treats the first extracted row of each selected page as its header. Review every cell. Merged cells, captions and complex tables may require cleanup or fail clearly.
Yes. Enter page numbers/ranges such as 2,1-3 in the desired order, or leave blank for all within the limit. Exported pages are separated with Markdown horizontal rules.
Images are not exported. Extracted text is escaped for Markdown, and the browser preview displays literal source rather than executing HTML. No OCR, AI or external extraction service is used.
Inputs stay in browser memory. PDFs are limited to 25 MiB and 200 source pages; extraction to 30 selected pages and 200,000 UTF-8 output bytes. Reset/cancel discards pending results and downloads require an explicit action.
Extract text-based PDF pages into downloadable Markdown with page selection, visible lists, optional headings and simple table mode.
Best for horizontally oriented single-column text. Rows follow visual top-to-bottom/left-to-right positions; gaps suggest paragraph breaks, and visible list markers become Markdown syntax. Default output does not invent headings. Opt-in heading inference uses conservative size cues and leaves uncertain short pages unstructured. Table mode reuses simple aligned-table extraction and treats each selected page’s first row as a header; review every cell. Complex layouts can require cleanup. Text syntax is escaped; images are not exported. No OCR/AI. Limits: 25 MiB PDF, 200 source pages, 30 selected pages and 200,000 UTF-8 output bytes.
Text-based, horizontally oriented, single-column pages work best. Scans without text need OCR elsewhere. Complex columns and layouts can produce reading-order ambiguities and require manual cleanup.
Visible list markers become Markdown list syntax. Heading inference is opt-in and relies on clear font-size differences; uncertain short pages stay plain text. Default extraction avoids inventing headings.
Optional table mode reuses the aligned text-table extractor and treats the first extracted row of each selected page as its header. Review every cell. Merged cells, captions and complex tables may require cleanup or fail clearly.
Yes. Enter page numbers/ranges such as 2,1-3 in the desired order, or leave blank for all within the limit. Exported pages are separated with Markdown horizontal rules.
Images are not exported. Extracted text is escaped for Markdown, and the browser preview displays literal source rather than executing HTML. No OCR, AI or external extraction service is used.
Inputs stay in browser memory. PDFs are limited to 25 MiB and 200 source pages; extraction to 30 selected pages and 200,000 UTF-8 output bytes. Reset/cancel discards pending results and downloads require an explicit action.