Set up PaddleOCR Skills
Install and verify the two skills from the linked repository. Treat this setup
skill as an installer guide; run OCR with the installed target skills.
Use this setup skill when the request mentions PaddleOCR installation or Agent
Skills, or asks for 图片转文字 / 截图识字 / 扫描件 OCR / PDF 转 Markdown /
表格提取 / 公式识别 / 版面分析. It installs:
paddleocr-text-recognition for screenshots, photos, scans, PDFs, CJK text,
line-level OCR, and optional bounding boxes.
paddleocr-doc-parsing for structured Markdown/JSON, tables, formulas,
figures, multi-column layout, and correct reading order.
Install
Confirm the user wants a global, user-level installation.
Check Node.js/npx, Python 3.9+, and uv. Explain any missing prerequisite
before using its official installer. Do not use sudo or change unrelated
system settings without permission.
Run:
npx skills add Aidenwu0209/PaddleOCR-Skills --skill '*' -g -y
Run npx skills list -g --json and confirm both
paddleocr-text-recognition and paddleocr-doc-parsing, including their
actual install paths.
Configure safely
- Use https://www.paddleocr.com for the official API and token flow.
- Ask the user to provide the OCR or document-parsing endpoint only when that
target skill needs it.
- Never invent, print, or store
PADDLEOCR_ACCESS_TOKEN in chat, command
history, source files, or logs.
- Report incomplete configuration instead of claiming the OCR service works.
Finish
Report prerequisite versions, commands executed, installed skill names and
paths, and any configuration still required from the user.
1---
2name: paddleocr-skills-setup
3description: >-
4 Install and configure two PaddleOCR Agent Skills for text recognition and structured document parsing in Codex, Claude Code, GitHub Copilot, Cursor, OpenCode, OpenClaw, and other compatible agents. Use for OCR and image-to-text from screenshots, photos, scans, and PDFs; Chinese/CJK text and bounding boxes; PDF-to-Markdown/JSON; tables, formulas, layout, and reading order; or endpoint, token, installation, and troubleshooting help.
5license: MIT-0
6metadata:
7 source: https://skillmd.com/skills/aidenwu0209/paddleocr-skills-setup
8 author: Aidenwu0209
9 repository: https://github.com/Aidenwu0209/PaddleOCR-Skills
10 version: "1.0.2"
11 openclaw:
12 emoji: "🧰"
13 homepage: https://github.com/Aidenwu0209/PaddleOCR-Skills
14---
15
16# Set up PaddleOCR Skills
17
18Install and verify the two skills from the linked repository. Treat this setup
19skill as an installer guide; run OCR with the installed target skills.
20
21Use this setup skill when the request mentions PaddleOCR installation or Agent
22Skills, or asks for **图片转文字 / 截图识字 / 扫描件 OCR / PDF 转 Markdown /
23表格提取 / 公式识别 / 版面分析**. It installs:
24
25- `paddleocr-text-recognition` for screenshots, photos, scans, PDFs, CJK text,
26 line-level OCR, and optional bounding boxes.
27- `paddleocr-doc-parsing` for structured Markdown/JSON, tables, formulas,
28 figures, multi-column layout, and correct reading order.
29
30## Install
31
321. Confirm the user wants a global, user-level installation.
332. Check Node.js/npx, Python 3.9+, and `uv`. Explain any missing prerequisite
34 before using its official installer. Do not use `sudo` or change unrelated
35 system settings without permission.
363. Run:
37
38 ```bash
39 npx skills add Aidenwu0209/PaddleOCR-Skills --skill '*' -g -y
40 ```
41
424. Run `npx skills list -g --json` and confirm both
43 `paddleocr-text-recognition` and `paddleocr-doc-parsing`, including their
44 actual install paths.
45
46## Configure safely
47
48- Use https://www.paddleocr.com for the official API and token flow.
49- Ask the user to provide the OCR or document-parsing endpoint only when that
50 target skill needs it.
51- Never invent, print, or store `PADDLEOCR_ACCESS_TOKEN` in chat, command
52 history, source files, or logs.
53- Report incomplete configuration instead of claiming the OCR service works.
54
55## Finish
56
57Report prerequisite versions, commands executed, installed skill names and
58paths, and any configuration still required from the user.