Deepseek Ocr

This is a Claude Code skill that wraps DeepSeek's OCR vision model for extracting text from images, PDFs, and documents. It's built for production use with vLLM inference (hitting around 2500 tokens/s on an A100) and supports batch processing, dynamic resolutions, and multiple prompt modes including document-to-markdown conversion, free OCR, and grounded text location. The setup is a bit heavy with specific CUDA and PyTorch requirements, plus you'll need to grab a vLLM wheel from their releases, but once running it handles things like table detection with whitelisted tokens and context optical compression for longer documents. Good option if you're processing documents at scale and want structured markdown output rather than just raw text strings.

skills.sh installs: 1.5K
GitHub stars: 61
last updated: Jul 14, 2026

View on GitHub

Install

$ npx -y skills add aradotso/trending-skills --skill deepseek-ocr --agent claude-code

Installs into .claude/skills of the current project.

Files

DeepSeek-OCR

Skill by ara.so — Daily 2026 Skills collection.

DeepSeek-OCR is a vision-language model for Optical Character Recognition with "Contexts Optical Compression." It supports native and dynamic resolutions, multiple prompt modes (document-to-markdown, free OCR, figure parsing, grounding), and can be run via vLLM (high-throughput) or HuggingFace Transformers. It processes images and PDFs, outputting structured text or markdown.

Installation

Prerequisites

CUDA 11.8+, PyTorch 2.6.0
Python 3.12.9 (via conda recommended)

Setup

git clone https://github.com/deepseek-ai/DeepSeek-OCR.git
cd DeepSeek-OCR

conda create -n deepseek-ocr python=3.12.9 -y
conda activate deepseek-ocr

# Install PyTorch with CUDA 11.8
pip install torch==2.6.0 torchvision==0.21.0 torchaudio==2.6.0 \
  --index-url https://download.pytorch.org/whl/cu118

# Download vllm-0.8.5 whl from https://github.com/vllm-project/vllm/releases/tag/v0.8.5
pip install vllm-0.8.5+cu118-cp38-abi3-manylinux1_x86_64.whl

pip install -r requirements.txt
pip install flash-attn==2.7.3 --no-build-isolation