Editorial recipe — not benchmarked — reviewed September 3, 2026
Choose a document parser without losing decision-critical evidence
For researchers and operators converting authorized pdfs or office files into model-readable text.
Run this recipe
Work through it here
Progress stays in this browser. The downloaded Markdown kit works in any notes app or repository.
0/5 complete
Is this recipe useful?
Procedure
- Select a small authorized fixture containing headings, tables, footnotes, citations, page references, and one deliberately difficult layout; record source hashes.
- Run pinned MarkItDown and Docling versions separately in restricted containers and retain their raw Markdown, structured output, warnings, and elapsed time.
- Score each output against the originals for ordering, table cells, references, omitted text, OCR errors, and usable source locations before involving a model.
- Give the stronger extraction to a local Qwen model and request an Obsidian claim ledger whose material claims point back to verified source locations.
- Have a person reopen every decision-changing number, date, obligation, and citation, then publish the parser decision with failures and document types where it does not hold.
Acceptance artifact
A hashed document fixture, two raw extraction sets, fidelity scorecard, source-linked claim ledger, and reviewed parser decision
Do not use it blindly
Neither parser is a security boundary; MarkItDown MCP can read files available to its process, so restrict mounts, network access, and input provenance.