docling-project/docling

Choose this fork only if its CSV and runtime-control additions are worth accepting an old, heavily diverged codebase. For most new adopters, upstream is the safer choice because it is far more current and actively maintained.

Prefer upstream unless you specifically need this exact older snapshot. This fork adds no visible functionality, and its main tradeoff is falling behind active upstream improvements.

Prefer this fork only if its local test corpus or workflow customizations are the asset you need. For most adopters, upstream is the better choice because this fork is stale, materially diverged, and likely missing recent bugfixes and capabilities.

Prefer upstream unless you specifically need the fork’s older CLI/PDF behavior or its lighter maintenance footprint. This fork looks materially behind and is a poor choice for users who want current Docling features, fixes, and integrations.

Choose this fork only if you want an older, debug-oriented Docling snapshot and can accept being materially behind upstream. If you want current format support, stability fixes, or ongoing maintenance, upstream is the better default.

Prefer upstream unless you specifically need this exact January 2026 snapshot. This fork adds no visible capabilities and is materially behind on current fixes and features, so it is a poor default choice for production adopters.

Choose upstream unless you specifically need the fork's older MLX/VLM, OCR-plugin, or DoclingParseV4 work. This fork is materially stale and carries substantial compatibility and maintenance risk.

Prefer this fork only if its custom OCR/artifact/image workflows are the goal. If you want the current Docling feature set, active maintenance, and upstream compatibility, upstream is the safer choice.

Choose this fork only if you specifically need its older behavior plus the small set of local fixes; otherwise upstream is the better default because this fork is stale and materially behind.