How this shortlist was built
The shortlist covers lightweight converters, structured open-source parsers, cloud document APIs, and hosted ingestion layers. We prioritize readable hierarchy, table and equation behavior, traceability, and operational fit rather than the mere presence of Markdown output.
Every capability statement is bounded to linked first-party documentation reviewed on the page's update date. Pricing, quotas, preview status, regions, model versions, and commercial terms must be checked again during procurement.
- No paid placement or affiliate ranking.
- No universal accuracy winner without a shared corpus.
- No star rating, market-share, or customer-count inference.
- Each candidate is assigned a best-fit operating boundary rather than a generic score.
The shortlist and where each tool fits
Treat this as a map of architectures, not a podium. A lightweight local converter, a cloud OCR processor, a model-led parser, and a tenant-aware ingest service solve different amounts of the stack.
The useful question is which layer your team is prepared to build and operate after the parser returns.
- DocParse — best for a hosted API that pairs Markdown with DocIR, routing, lifecycle, and tenant controls. Digital PDFs can use a deterministic text-layer route while scans and complex layouts remain explicit options.
- Marker — best for customizable local conversion with OCR and optional LLM correction. The project outputs Markdown, JSON, HTML, or chunks and documents fast, balanced, OCR, and no-OCR modes.
- MinerU — best for self-hosted complex-document conversion and scalable inference. The project targets PDF and Office parsing into Markdown and JSON with router and multi-GPU options.
- Docling — best for a structured local conversion toolkit and rich document model. Docling supports multiple formats and exposes a document converter plus structured outputs.
- Microsoft MarkItDown — best for lightweight Python and CLI file-to-Markdown conversion. Optional extras activate formats such as PDF, DOCX, PPTX, and XLSX.
- Mistral OCR — best for a focused managed OCR API with document structure. The official document-processing API returns page content and document elements for OCR workflows.
- Adobe PDF to Markdown — best for teams already using Adobe PDF Services. Adobe documents a dedicated asynchronous PDF-to-Markdown operation alongside PDF Extract.
- Azure Document Intelligence — best for Azure-native layout Markdown. The layout API returns Markdown, sections, figures, and HTML table representations.
Choose by requirements that survive a demo
Write requirements in terms of inputs, output evidence, deployment, product operations, and acceptable failure behavior. Avoid choosing by a single polished sample or an aggregate benchmark that does not match the production distribution.
Separate must-have constraints from quality dimensions you can benchmark. Region, license, language, file type, and deployment can eliminate a candidate before accuracy testing begins.
- Born-digital versus scanned PDFs
- Multi-column reading order
- Merged tables and figures
- Equations and inline math
- Page and block traceability
- Local library versus hosted asynchronous API
Turn the shortlist into a production decision
Create a frozen evaluation corpus, version the expected assertions, and run each candidate with documented settings. Keep source bytes and sensitive material inside the authorization boundary established for the evaluation.
Publish results by cohort with failures and uncertainty visible. Then test the downstream pipeline: a structurally faithful document can still produce poor chunks, and a readable Markdown file can still lack the page evidence required for citations or deletion.
- Compare rendered meaning, not character count
- Test tables across page boundaries
- Retain page identifiers
- Inspect repeated header removal
- Measure OCR-trigger rate
- Version outputs and parser configuration
Frequently asked questions
Questions teams ask before building
Is the first item the most accurate tool?
No. Items are organized as an architecture shortlist, not a universal accuracy ranking. Accuracy must be measured on your documents with fixed versions, settings, and acceptance rules.
Can pricing alone decide the shortlist?
Page price is only one component. Include retries, model calls, storage, queue and webhook infrastructure, human review, migration work, monitoring, and the engineering cost of self-hosting.
How often should this decision be reviewed?
Recheck official documentation before procurement and rerun the frozen evaluation when a parser, model, API version, document distribution, or downstream acceptance rule materially changes.