Our take
- Activity
- 29
- Community
- 17
- Issues
- 21
- Pull requests
- 61
-145 stars in the last 30 days
- Fewer contributors than 82% of the projects we track
Project history
Daily GitHub snapshots since Oct 7, 2026. History covers stars and forks; health signals are not tracked historically.
Stars
1,545Forks
363Change since Nov 16, 2026: 0 stars·0 forks
Overview
docling-api wraps Docling in a deployable REST service for converting formats such as PDF, DOCX, PPTX, HTML, CSV, and images into Markdown. It addresses the gap between a document-parsing library and a service that applications can call over an API.
The project combines FastAPI with Celery and Redis for queued work, including asynchronous and batch conversion. It is aimed at developers and operators building document-ingestion workflows who want to run and scale the conversion backend themselves.
Key Features
- Converts PDFs, office documents, HTML, CSV, images, AsciiDoc, and Markdown.
- Extracts text and tables, with an option to represent tables as images.
- Supports multilingual OCR and image-resolution configuration.
- Offers synchronous single-document and batch endpoints.
- Offers asynchronous conversion jobs with status tracking.
- Uses Celery and Redis for distributed processing, with Flower for task monitoring.
- Supports CPU-only and GPU-enabled deployments.
Use Cases
- A document-ingestion service for a search or knowledge-base application that needs Markdown output.
- Batch processing of mixed-format files by teams handling large collections of documents.
- An OCR pipeline for scanned pages or image-based documents in supported languages.
- A backend for applications that need to submit conversions asynchronously and track job status.
Getting Started
Clone the repository and start the CPU-mode Docker services:
git clone https://github.com/drmingler/docling-api.git
cd docling-api
docker compose up --build --scale celery_worker=1
See the README and Docker guide for configuration, GPU setup, and API examples.
Alternatives
- text-extract-api: text-extract-api handles PDF, Word, and PowerPoint extraction, with OCR, LLM-based PII removal, and structured JSON output alongside Markdown.
- Ollama-OCR: Ollama-OCR extracts text and structured content from images and PDFs using Ollama vision-language models, as a Python package or Streamlit app.
| Project | Language | License | Stars | Status |
|---|---|---|---|---|
| docling-api | Python | MIT | 1.7k | Not checked yet |
| text-extract-api | Python | MIT | 3.2k | Quiet |
| Ollama-OCR | Jupyter Notebook | MIT | 2.8k | Inactive |
Considerations
- Running the service requires operating its API, Redis, and Celery components; it is not a hosted conversion endpoint.
- The README lists Python 3.12 or higher, Poetry, and Redis for local setup. Docker is another documented deployment path.
- GPU support is optional, but the README recommends it for production and large-document performance. GPU availability and deployment add operational requirements.
- Conversion quality and throughput can depend on document complexity, chosen processing mode, and worker capacity. The repository does not provide benchmark figures in the supplied information.
Found this useful?
Share it with someone who would like docling-api.