Automate patent invention disclosures with artificial intelligence

What you get

• Structured patent text generation: Creating the technical field, background, abstract, and detailed description sections of an invention in accordance with standard patent norms.
• Independent and dependent claim tree: Automatically formulating hierarchical patent claim lists that maximize the scope of legal protection.
• Prior art gap analysis: Clearly highlighting the technical differences and the inventive step between existing technologies and the invention.
• Accelerating collaboration with patent attorneys: Saving time and money by transforming engineers' drafts into ready-to-use, organized technical documents for patent attorneys.
• Multilingual patent terminology support: Compliance with English, Turkish, and international patent office (WIPO, EPO, USPTO) terminology.

Installation

git clone https://github.com/handsomestWei/patent-disclosure-skill.git
cd patent-disclosure-skill
pip install -r requirements.txt

Running it

python run_skill.py --input bulus_taslagi.txt --output patent_disclosure.md

Technical architecture and working principle

• Technical Invention Decomposition Engine: Identifies key inputs, outputs, and methodology in software, hardware, or chemical process descriptions.
• Claim Syntax Validator: A legal language analyzer that checks for vague expressions and formal errors in claims.
• Template and Markdown Export: Saving the document in a standard sectioned Markdown format for use in official patent applications.

Patent analysis workflows and claim preparation

• Converting Software Algorithms into Patentable Formats: Deriving method and system descriptions acceptable to patent authorities from code and architectural diagrams.
• Defense Against Office Actions: Drafting responses that list the distinctive features of an invention in response to objections from patent examiners.
• Intellectual Property Portfolio Audit: Early mapping of the inventive steps within in-house technological projects that hold patent potential.

If you don't write code

Frequently asked questions

• Does this tool replace a registered patent attorney? No. Patent Disclosure Skill is a preliminary preparation and productivity tool that helps engineers draft their inventions and prepare them for attorneys; legal filings must be handled by an attorney.
• Which LLM models does it work with? It can be configured to work with Claude 3.5 Sonnet, GPT-4o, or local open-weight models (Qwen, Llama 3).
• Will my confidential technical secrets leak to the internet? When running with a local LLM (Ollama or vLLM), all patent analyses are performed entirely on your local computer, and no data leaves your machine.
• Can it interpret patent drawings and flowcharts? When multimodal models are connected, it can also analyze system architecture and block diagram images and transcribe them into text.

Related dictionary terms

Links

• GitHub repository →
• Read in Turkish →