patent-download

Download patent PDFs and query patent metadata from Google Patents and Chinese patent platforms.

636|91|Updated Jan 7, 2026
One-click install
npx skills add https://github.com/cat-xierluo/legal-skills --skill patent-download
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: patent-download
Source: https://github.com/cat-xierluo/legal-skills/tree/main/skills/patent-download
Command: npx skills add https://github.com/cat-xierluo/legal-skills --skill patent-download

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires patent-downloader, playwright, requests, and includes scripts (resource) and references (resource) components.

What problem does it solve?

Manually searching patent databases and downloading full-text PDFs one by one is slow and error-prone, especially when users only have an application number while platforms index by publication number. This Skill automates patent PDF retrieval and metadata lookup across multiple platforms.

Core Features & Use Cases

  • Multi-platform patent download: Retrieve patent PDFs via Google Patents (free, no login, recommended), with fallback channels including uyanip, PatentStar, GPIC, PSS, and the CNIPA epub system.
  • Application number to publication number handling: Automatically parses Chinese application numbers (e.g., 202421964517.8) and attempts to match publication numbers (e.g., CN223081266U), with differentiated failure hints by patent type.
  • Batch download and info-only queries: Download multiple patents in one command or query title, assignee, and publication date without downloading.
  • Use Case: A patent attorney needs the granted specification of a utility model for litigation. They run the CLI with the application number, and the Skill resolves it to CN223081266U and saves the PDF to the chosen output directory.

Quick Start

Ask the AI to download the full-text PDF of patent CN223081266U from Google Patents into your Downloads folder.

Frequently Asked Questions about patent-download

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I download a Chinese patent PDF from Google Patents?

Run python cli.py google followed by the publication number, such as CN223081266U, after installing the patent-downloader package. Google Patents is free and requires no login, and you can pass multiple numbers for batch download.

What is the difference between a patent application number and a publication number?

An application number like 202421964517.8 is assigned at filing, while a publication number like CN223081266U is assigned at publication. Google Patents indexes by publication number, and there is no mathematical mapping between the two.

Why does Google Patents fail to find my patent number?

Lookup fails when you enter an application number that cannot be auto-matched, a mistyped publication number, or a design patent, which Google Patents covers poorly. The tool prints type-specific hints suggesting alternatives like uyanip or the CNIPA epub system.

Can I query patent information without downloading the PDF?

Yes, use the --info flag, for example python cli.py google CN223081266U --info. This prints the title, assignee, publication date, and abstract using only the Python standard library with no extra dependencies.

How are platform accounts and passwords configured?

Credentials are read from environment variables named PATENT_<PLATFORM>_USERNAME and PATENT_<PLATFORM>_PASSWORD, or from a local config/.env file that is git-ignored. A self-check script verifies the .env file was not accidentally committed to git.

Which patent download platforms actually work in this tool?

Google Patents is the verified primary channel and uyanip browser automation is a working backup. PatentStar's API is defunct, and the GPIC, PSS, and epub modules are experimental stubs that do not save PDF files.