microsoft/markitdown

Prefer upstream unless you specifically want a frozen baseline. This fork adds nothing observable over upstream and is behind on recent fixes and workflow improvements, so it is mainly useful as a starting point for private customization.

Choose this fork if your priority is a leaner dependency stack and you do not need Magika-related behavior. Prefer upstream if you want the newest fixes, broader compatibility, or the full optional-feature set.

Choose upstream unless you explicitly need an older frozen baseline. This fork adds no visible capabilities and is materially behind on features, performance fixes, and integration work.

Prefer upstream unless you specifically need this exact pinned revision; this fork adds no new capability and is already behind upstream bugfix/performance work.

Prefer upstream unless you specifically need a frozen 2024-12 snapshot. This fork adds no visible capabilities and is materially behind on maintenance and feature work.

Prefer this fork only if you want a stable, mostly upstream baseline to pin against. If you want the newest PDF/OCR fixes and ongoing maintenance, upstream is the better choice.

Adopt this fork only if you specifically want a frozen base for your own work; otherwise, upstream is a better default because it is newer and has added extraction and reliability improvements.

Choose this fork if you want a more opinionated LLM ingestion pipeline with chunking, image-url output, and stronger PDF/image handling. Stick with upstream if you need the broader Markdown converter ecosystem, embedding support, or maximum compatibility with MarkItDown's original APIs and plugin surface.

Prefer upstream unless you specifically need this exact snapshot. This fork offers no visible added capability and is materially behind recent upstream maintenance, so it is best treated as a stale baseline rather than a better distribution.