Elysia Tools
Navigation
Workflow Playbook
Recover Markdown from PDFs or images, export Markdown to PDF, HTML, slides, reports, plain text, JSON, or CSV, and verify the result for the next channel.
Hubs
Markdown conversion is safest when the source and target are explicit. A README, pasted Markdown file, scanned PDF, image, and text-layer PDF do not need the same handling. Start by deciding whether you are recovering editable Markdown, publishing a readable artifact, or extracting structure for another system.
Use pdf-to-markdown for text-layer PDFs, scanned-pdf-ocr-to-markdown for scanned pages, and ai-image-to-markdown when the source is an image. For publishing, choose between markdown-to-pdf-converter, markdown-to-pdf-theme-pack, markdown-report-bundler, md-slide-deck-to-pdf, or markdown-to-html based on the final channel rather than running every exporter.
Human-facing output should be checked for headings, code blocks, tables, links, page breaks, and styling. Machine-facing output from markdown-to-json or markdown-to-csv should be parsed and compared against representative fields or rows. A useful conversion record names the source, target, known OCR or formatting losses, and the downstream acceptance check.
Workflow playbook
Start by deciding whether the input is already Markdown, a PDF with text, a scanned PDF, an image, or a GitHub README. Recover editable Markdown only from the matching source branch, then inspect the first result before further export.
When the goal is readable text or a lightweight review artifact, remove syntax deliberately or convert to HTML while checking links, code fences, tables, and heading hierarchy.
Choose the PDF branch that matches the deliverable: a general Markdown PDF, a themed print-ready PDF, a multi-file report bundle, or a slide deck with one page per slide.
If the Markdown contains tables or structured sections for another system, export only the required machine format and compare representative fields or rows after parsing.
Choose the branch that matches the real source. Markdown publishing, scanned PDF OCR, image extraction, and README fetching have different evidence and cleanup requirements.
Use PDF, themed PDF, report bundles, slides, or HTML for human reading; use JSON or CSV only when the Markdown structure or tables need to feed another system.
Check headings, links, code fences, tables, page order, image-derived text, and metadata that downstream readers or parsers depend on.