dococr-skill

dococr-skill is a Codex skill for working with docOCR.

The docOCR project repository is riddleling/docOCR. docOCR is a local macOS OCR-to-Markdown tool. It can print OCR Markdown to stdout from the CLI, write .md files with -o, run a local HTTP server, and expose a JSON OCR API at /api/ocr.

Purpose

This skill gives Codex reusable context and workflows for docOCR-related tasks, including:

• Explaining docOCR CLI, server, and API usage.
• Running docOCR on images and returning Markdown from stdout or writing .md files.
• Debugging local API, multipart upload, port, or server startup issues.

Installation

Clone this repository to any local directory:

git clone https://github.com/riddleling/dococr-skill.git

Then link the skill into Codex's skills directory:

mkdir -p ~/.codex/skills
ln -s /path/to/dococr-skill ~/.codex/skills/dococr-skill

After installation, start a new Codex thread or restart Codex so it reloads available skills.

Usage

Explicitly invoke the skill in Codex with $dococr-skill:

Use $dococr-skill to explain how docOCR's /api/ocr endpoint works.

Use $dococr-skill to OCR /path/to/Desktop/01.png and return the Markdown printed by docOCR.

Use $dococr-skill to OCR /path/to/Desktop/01.png with docOCR -o and write the Markdown file next to the image.

Use $dococr-skill to OCR /path/to/Desktop/01.png through the local API at http://127.0.0.1:8080/api/ocr and return the Markdown text.

Use $dococr-skill to troubleshoot a failed docOCR upload to http://127.0.0.1:8080/api/ocr.

To run OCR for real, docOCR must be installed locally. See riddleling/docOCR for docOCR installation and usage details.

Directory Structure

dococr-skill/
├── SKILL.md
├── agents/
│   └── openai.yaml
├── README.md
└── .gitignore

SKILL.md contains the main skill instructions. agents/openai.yaml contains Codex UI metadata.