Docling CLI
Docling is a CLI for mixed document archives, bringing PDFs, office files and spreadsheets into a shared format for AI search.
- 70,176 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The CLI reads documents and writes result files; URL sources require network retrieval.
Install
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/doclingDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The CLI reads documents and writes result files; URL sources require network retrieval.
This is third-party code. Review the repository files before installing.
What it does
Docling is a CLI for converting PDFs, office files and other documents into structured content. Users supply a local path or URL and receive a Markdown file. The parser handles reading order, tables and page elements so the text can feed search or AI applications. OCR recognizes text in scans. A vision-language model (VLM) can process visually complex pages. The standalone `docling` command is separate from the Docling MCP server, which exposes the engine to agents through the protocol.
Who it is for. Knowledge-base developers preparing complex-layout documents for search and source-based answers.
Good fit when
- A PDF has columns and tables that need structured text extraction.
- A team normalizes DOCX, PPTX and other inputs into a shared representation.
- A scanned document needs OCR before indexing.
- An analyst compares the default PDF pipeline with a visual model on a difficult page.
Not a fit when
- An agent needs an MCP tool but only the CLI Python package has been installed.
- An editor expects every complex formula to be recovered without reviewing output.
- The system relies on author and reference metadata extraction currently listed under Coming soon.
Example request
Convert a test PDF with the Docling CLI and show reading order and a Markdown table before indexing.Limitations
Python 3.10 or later is required after support for 3.9 ended in version 2.70.0. MIT covers the codebase, while model usage follows the licenses of the original model packages. Extraction of title, author, reference and language metadata remains under Coming soon. An air-gapped workflow needs its local environment and models prepared in advance; supplying an external document URL still requires network access to retrieve that source.
How to disable. Stop conversion and remove the command from the pipeline. Saved Markdown and JSON do not require a running Docling process; disable any separate MCP server in its client.
FAQ
How do I install the CLI?
Install `pip install docling` in a Python 3.10 or later environment. Pass a small PDF to `docling` and inspect its Markdown before processing the archive.
Where is the Markdown output?
`docling https://arxiv.org/pdf/2206.01062` creates a `.md` file in the current directory. Open it and compare table structure and paragraph order with the PDF.
How do I try a visual model?
The README gives `docling --pipeline vlm --vlm-model granite_docling https://arxiv.org/pdf/2206.01062`. This selects a separate VLM pipeline; compare a sample page before processing a large set.
Is this the same tool as Docling MCP?
The CLI runs directly in a shell. The MCP server is a separate agent adapter; use the `docling-mcp` card and setup for that integration.
Which office formats are supported?
The README lists DOCX, PPTX and XLSX alongside ODT, ODS and ODP. Try a small file of the required type and review Markdown before batch processing.
Discussion
No questions or reviews yet.
Related
CLI toolsAn MCP server built on Docling: converts PDFs into structured documents, builds new ones and feeds them to RAG
A Firecrawl skill and CLI that lets your agent convert Word, PowerPoint, Excel, OpenDocument, RTF, EPUB, CSV and PDF to Markdown on your own machine
Doc-Ops MCP converts DOCX, Markdown, HTML and PDF between formats and stamps a watermark or QR code onto finished PDFs.
A local CLI extracts text from documents and scans, searches archives and prepares source material for an agent’s answer.
An MCP server on top of pandoc: the agent converts Markdown, HTML, DOCX, LaTeX, EPUB, ODT and other formats into one another
A skill plus the docx program: the agent fills forms, edits tables and reviews Word files with tracked changes
README in short
Results use the shared DoclingDocument structure, with Markdown and JSON exports. Scripts can invoke the CLI; applications can use the Python API. MCP access is supplied by a separate server.