Docling

Health v2 · maintenance only

Adoption not part of the score

68703 stars · 5048 forks observed · 2026-10-12

What it is AI-extracted, prompt v1, taxonomy v1, 2026-09-21, confidence 0.98

Docling is a Python SDK and CLI that parses a wide range of document formats into a unified representation. It also provides a service mode for API‑based document processing and integrations with LLM frameworks.

Use cases

• Convert PDFs, DOCX, and HTML files into clean Markdown for downstream LLM pipelines.
• Extract tables, code snippets, and formulas from scanned PDFs using OCR for data analysis.
• Prepare legal or financial documents for retrieval‑augmented generation (RAG) applications.
• Transcribe audio files and embed the transcript into a searchable document structure.
• Run Docling as a microservice to batch‑process documents in an air‑gapped environment.

When to choose

• You need a single tool that handles many file types (PDF, DOCX, PPTX, images, audio) and outputs structured JSON or Markdown.
• Your workflow requires local processing for privacy or compliance reasons.
• You want out‑of‑the‑box integrations with LangChain, LlamaIndex, or other LLM toolkits.

When to avoid

• You require a heavyweight OCR engine with custom model training beyond the built‑in support.
• Your project is limited to non‑Python environments without a REST API wrapper.
• You need real‑time streaming document processing at massive scale; a dedicated enterprise service may be more appropriate.

Facets

library · maturity active

parser cli http-server artificial-intelligence data-science python cli document-conversion pdf-parsing rag llm-integration markdown-export ocr audio-transcription natural-language-processing web-server

Member repositories

For agents

markdown · JSON · MCP: product_card(name="docling-project/docling")