Docling Serve

Docling Serve is an open-source toolkit from IBM Research that runs Docling as an API service, converting documents (PDF, DOCX, images, HTML) into structured…

  • AI
  • Productivity
  • Tools & Utilities
Docling Serve is an open-source toolkit from IBM Research that runs Docling as an API service, converting documents (PDF, DOCX, images, HTML) into structured Markdown or JSON for RAG and local document processing. It offers a REST API with an optional Web UI (default port 5001), multi-format parsing with layout understanding and table extraction, and runs locally so your data stays private — a good fit for homelab RAG pipelines.

Platform availability

Available on 1 of 10 platforms

  • Unraid
  • TrueNAS (not listed)
  • Umbrel (not listed)
  • ZimaOS (not listed)
  • Proxmox (not listed)
  • Helm (not listed)
  • Mac (not listed)
  • Windows (not listed)
  • Linux (not listed)
  • Any server (not listed)

Health score?A 0-100 score computed nightly from four factors: maintained (40%: recent commits and steady activity), popular (25%: GitHub stars, log scale), easy to install (20%: platforms, install notes, setup guides, Docker image) and light to run (15%: minimum RAM, ARM support). When the source code has been checked, a fifth factor, security check (20%: automated scan for committed secrets, unsafe desktop-app settings, known-vulnerable dependencies and project hygiene; any serious finding caps it at 20), is added. Factors without data are left out and the rest are rescaled.

55/100 Fair

Maintained
100last commit 25 days ago
Popular
661,914 GitHub stars
Easy to install
351 install platform, install notes, Docker image
Light to run
108 GB minimum RAM
Security check
54 serious findings, 23 minor, CI, tests

Availability: platform listings confirmed today, links checked 22 days ago

GitHub stars
1,914
Open issues
181
Last commit
2026-10-27
Latest release
v1.38.2
License
MIT
Activity
steady

Checked 7 days ago - source: GitHub (docling-project/docling-serve)

Resources & Compatibility

Minimum requirements: 8 GB RAM, 5 GB disk.

Checked 54 days ago - source: Docker Hub image tags and published docs

First-install notes

  • Install Docling Serve and enable its UI Use pip (the Python package installer) to install 'docling-serve[ui]', then run docling-serve run --enable-ui. Without those extras the web interface won't be available, and the server binds to 0.0.0.0 (all network interfaces), so allow it only from devices you trust. source
  • Use Docker Compose for an NVIDIA GPU If you are using Docker Compose (the tool that starts containers) with an NVIDIA GPU, run docker compose -f compose-nvidia.yaml up -d after installing the NVIDIA drivers and toolkit (the software that lets your computer use the GPU). Without those installed on your computer, the container cannot use your GPU. source
  • Set up Ollama for Granite-Docling If you use the Granite-Docling example, first install and start Ollama (the local model runner), then run ollama pull ibm/granite-docling:258m. If Ollama is not running or the model is missing, the example skips and that pipeline will not process documents. source
  • First-run model download On a fresh install the models directory is empty and Docling must download RapidOCR and other model artifacts on first boot. Set DOCLING_SERVE_ENABLE_REMOTE_SERVICES to true for the very first start so downloads can reach the upstream model hosts (e.g. modelscope.cn). source

Fetched 47 days ago - each note links to its source

Alternatives (1)

  • Markitdown Online - Markitdown Online quickly converts your documents (including DOCX, HTML, and more) into clean and structured Markdown format.

Checked 54 days ago - source: AlternativeTo

Common questions

How do I enable the optional web UI for Docling Serve?

Install the UI extra, for example 'docling-serve[ui]', and run 'docling-serve run --enable-ui'. Without those extras the web interface is not available.

What should I know about the server's default network exposure?

The server binds to 0.0.0.0, which is all network interfaces, so allow it only from devices you trust.

How do I use an NVIDIA GPU with Docker Compose?

Install the NVIDIA drivers and toolkit on the host, then run 'docker compose -f compose-nvidia.yaml up -d'. Without that host setup the container cannot use the GPU.

Does Docling Serve include Kubernetes or OpenShift deployment guidance?

Yes, guidance covers container orchestration patterns, persistent storage configuration, networking setup, and scaling considerations.

Answers sourced from deepwiki.com, docling-project.github.io

Community

Collected 44 days ago - public community threads