Docling
Version 2.131.0
Maintained by LF AI & Data Foundation (hosted project)278, IBM Research (originating team)26
Docling is an open-source Python toolkit that converts documents such as PDF, DOCX, PPTX, XLSX, HTML, images, audio, and other formats into a structured DoclingDocument that can be exported as Markdown, HTML, or JSON for generative AI applications. It includes PDF layout, reading-order, and table-structure analysis, OCR, and optional vision-language model pipelines. The project began at IBM Research in Zurich and has been hosted by the LF AI & Data Foundation since 2025.2678
- Repository: Repository (external site: github.com)
- Documentation: Documentation (external site: docling-project.github.io)
- License: LICENSE (external site: raw.githubusercontent.com)
- Paper: Docling Technical Report (arXiv 2408.09869) (external site: arxiv.org)
- Release notes: Release v2.131.0 (external site: github.com)
Availability and license
Overall availability
Source code on GitHub under the MIT License; installable from PyPI with pip or uv, with optional extras for OCR engines, ASR, and VLM pipelines.134
Availability is separate from permission: read the license before using or redistributing.
The README states the codebase is MIT-licensed and that individual models used by Docling carry the licenses of their original packages.2
Public materials checklist
| Item | Status | Notes and evidence |
|---|---|---|
| Source codeIs the source code publicly readable? | Public | 1 |
| DocumentationIs user documentation published? | Public | Documentation covers installation, usage, concepts, examples, and integrations; a technical report describes the architecture and models.26 |
| InstallationAre installation instructions or packages publicly available? | Public | Documented installation with pip or uv, with optional extras such as easyocr, rapidocr, tesserocr, asr, and vlm.4 |
| Supported platformsAre supported operating systems or hardware documented? | Public | Documented as working on macOS, Linux, and Windows on x86_64 and arm64; the README states Python 3.10 or higher is required since version 2.70.0.42 |
| Release statusAre versioned releases published? | Public | Versioned releases are published on GitHub; v2.131.0 was published on 2026-09-29.5 |
What it is useful for
Preparing documents for retrieval-augmented generation and agent workflows, with local execution for sensitive or air-gapped data, integrations with frameworks such as LangChain, LlamaIndex, CrewAI, and Haystack, a command-line tool, an MCP server, and an API server (docling-serve).2
Run and use notes
- The installation guide documents a CPU-only PyTorch install on Linux via the PyTorch CPU wheel index, and a mac_intel extra for Intel Macs.4
Organization context
U.S. eligibility
Eligible · basis: U.S.-governed project
Docling is hosted as a project of the LF AI & Data Foundation, which describes itself as an umbrella foundation of the Linux Foundation. The foundation's project page states that IBM donated Docling as an Incubation-stage project in March 2025 (its induction was announced in April 2025) and that it became a Graduate-tier project in August 2026. The Linux Foundation's privacy policy gives a mailing address in San Francisco, California. The project was started by IBM Research's Zurich team; IBM is recorded as the originating organization, and eligibility rests on the U.S.-based foundation hosting the project, not on where its developers work.2789
Sources
This listing is not an endorsement, a safety assessment, or a federal approval.