Ai

Docling 2.94.0: Tectonic TikZ Rendering and Enhanced Chunking Options

Docling 2.94.0: Tectonic TikZ Rendering and Enhanced Chunking Options

Docling just dropped version 2.94.0, and it’s a double feature. The AI document-processing toolkit now offers optional Tectonic TikZ rendering plus finer control over images in chunked output. Both additions target users who need precise LaTeX rendering or cleaner markdown extraction. Let’s dig into what changed and why it matters.

What Changed

The headline is LaTeX support. Docling already handled LaTeX documents, but TikZ figures—complex inline graphics—were often a pain point. With this release, you can optionally enable Tectonic TikZ rendering. That means Docling will leverage the Tectonic engine to properly render TikZ code into images. It’s optional, so if you don’t need it, nothing changes. But if you work with papers or reports full of diagrams, this is a game-changer.

Second, BaseChunkerOptions got two new fields: image_placeholder and use_markdown_images. These let you control how images appear in chunked output. Set image_placeholder to keep a reference to the image without embedding it; use_markdown_images switches between inline base64 and standard markdown image syntax. Simple, but powerful for downstream apps that need to handle images differently.

Why It Matters

These aren’t flashy new features, but they solve real friction. TikZ rendering is a big deal for academics and engineers who live in LaTeX. Before, you’d need an external tool or messy workaround. Now it’s a flag away. The chunking options matter for anyone building RAG pipelines or data extraction tools—image handling is often the weakest link. Personally, I’ve seen projects stumble on image embedding in markdown chunks. This update gives you the knob to tune it.

Docling continues to polish its core. No bloat, just targeted fixes. If you’re processing documents at scale, v2.94.0 is worth grabbing.

Official Source: https://github.com/docling-project/docling/releases/tag/v2.94.0