# AI use cases for Docling (docling-project/docling) Repo: [docling-project/docling](https://github.com/docling-project/docling) | Use cases: 3 | Page: https://dedupedai.com/use-cases/docling-project-docling/ Docling parses documents for generative AI. It reads PDF, DOCX, PPTX, XLSX, HTML, EPUB, audio, video, email and more, with advanced PDF understanding of page layout, reading order, tables, code and formulas, and exports to Markdown, HTML, DocTags or lossless JSON. It runs locally, supports OCR and visual language models such as GraniteDocling, and plugs into LangChain, LlamaIndex, Crew AI, Haystack and MCP agents. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Convert a complex PDF with tables into clean Markdown: https://dedupedai.com/use-cases/docling-project-docling/convert-a-research-pdf-with-tables-to-markdown/ - Prepare a document folder for a RAG index: https://dedupedai.com/use-cases/docling-project-docling/prepare-documents-for-a-rag-index/ - Let your AI agent read documents through Docling: https://dedupedai.com/use-cases/docling-project-docling/give-your-agent-document-reading-via-mcp/ ## How to convert a complex PDF with tables into clean Markdown with Docling Repo: [docling-project/docling](https://github.com/docling-project/docling) | Difficulty: Beginner | Time: 15 minutes | Basis: From the README Install Docling and convert a paper or report, keeping layout, reading order and tables intact. **Who it is for:** A researcher who keeps getting jumbled columns and broken tables when extracting text from PDFs. **The problem:** Two-column layouts, tables and formulas come out scrambled with simple PDF text extraction. **The outcome:** A Markdown file in the current directory with the document's structure, tables and reading order preserved. ### Steps 1. Install with `pip install docling` 2. Convert a document from a URL or a local path with `docling https://arxiv.org/pdf/2206.01062` 3. Open the generated .md file in the current directory 4. Try the visual language model pipeline with `docling --pipeline vlm --vlm-model granite_docling https://arxiv.org/pdf/2206.01062` and compare ### Prompt to paste into your agent ```text Install docling-project/docling and convert to Markdown with the docling CLI. Then run it again with the granite_docling VLM pipeline and tell me which version handled the tables better. ``` ### Caveat The first run downloads models, and the VLM pipeline is much slower without a GPU. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/docling-project-docling/convert-a-research-pdf-with-tables-to-markdown/ ## How to prepare a document folder for a RAG index with Docling Repo: [docling-project/docling](https://github.com/docling-project/docling) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README Use the Python API to convert mixed documents into Markdown before chunking them into your retriever. **Who it is for:** A developer building a chat-with-your-docs app over a pile of PDFs, slides and spreadsheets. **The problem:** Garbage in, garbage out: badly parsed documents give a RAG system bad chunks and wrong answers. **The outcome:** Structured Markdown exports of every document, ready to chunk and embed in LangChain, LlamaIndex or Haystack. ### Steps 1. Install with `pip install docling` 2. In Python, create a `DocumentConverter()` and call `converter.convert(source)` with a local path or URL 3. Export each result with `result.document.export_to_markdown()` 4. Feed the Markdown into your framework; the README lists integrations with LangChain, LlamaIndex, Crew AI and Haystack 5. Keep sensitive files on your machine; Docling supports local execution for air-gapped environments ### Prompt to paste into your agent ```text Using docling-project/docling, write a Python script that converts every file in with DocumentConverter, saves the Markdown with export_to_markdown, and loads the results into for retrieval. Run it and report any files that failed. ``` ### Caveat The README points to the docs for chunking and framework examples, so plan to read the integrations page. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/docling-project-docling/prepare-documents-for-a-rag-index/ ## How to let your AI agent read documents through Docling with Docling Repo: [docling-project/docling](https://github.com/docling-project/docling) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Connect Docling to an agent through its MCP server or run it as a service with docling-serve. **Who it is for:** Someone using Claude Code or another MCP client who wants it to parse PDFs and office files properly. **The problem:** Agents often read PDFs as raw text and miss tables, figures and reading order. **The outcome:** An agent that can call Docling to parse documents, or a team service that converts files over an API. ### Steps 1. Install Docling with `pip install docling` to test conversions locally first 2. Follow the MCP server docs to connect Docling to your agent 3. For a shared service, follow the API server docs for docling-serve 4. Ask the agent to parse a document and check the structure it returns ### Prompt to paste into your agent ```text Set up docling-project/docling for my agent using its MCP server, following the Docling MCP docs. Then parse through it and summarise the tables it found. ``` ### Caveat The README only links to the MCP and API server docs; setup details live there, not in the README. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/docling-project-docling/give-your-agent-document-reading-via-mcp/