Public Prompt License (PPL) – prompt-native licensing for LLM prompts (shipfail.github.io)

A new licensing initiative called the Public Prompt License (PPL) has been proposed to address the unique challenges of licensing in the realm of AI and machine learning, particularly for large language model (LLM) prompts. Traditional licenses, such as MIT and AGPL, are not well-suited to the nuances of AI agents, creating ambiguity around what needs to be disclosed when these agents operate as services. The PPL offers a clearer structure by defining "Prompt Source" — which encompasses cognitive logic, personas, and tool descriptions — while intentionally excluding underlying infrastructure like inference engines and operating systems. This focus allows developers to share essential prompt logic without exposing sensitive operational details, addressing concerns many had with overreaching licenses like the Server Side Public License (SSPL). The significance of the PPL lies in its effort to foster a more collaborative environment for AI/ML developers, enabling them to construct and share innovative solutions without compromising their proprietary methods or secrets. The licensing framework provides three distinct variants—MIT, Apache, and Service—giving developers flexibility based on their project requirements. Currently in draft phase v0.1, the PPL seeks community feedback to refine its approach, aligning with Open Source principles while safeguarding both prompt integrity and user security in the rapidly evolving landscape of AI development.