unknown

Trace is an open-source, declarative specification standard designed to improve AI-assisted code generation and teach developers systems thinking. It organizes software specs into a structured folder format containing an overview with a global glossary, a stack file, domain files (with entities, events, rules, constraints, and externals), persona files (with access rules and user research), and flow files (using normalize/denormalize UX strategies). By writing clear, structured specs first, developers give LLMs precise context—reducing vague output—while also building architectural skills around domain boundaries, event-driven communication, and data ownership. The spec includes an agent skill with Author Mode (guided spec creation) and Implement Mode (phased code generation with review gates). The project is open source on GitHub.

Trace: Declarative Modeling for the AI Age

Tony

Trace is a markdown-based declarative modeling approach to spec writing for LLM consumption.

It's designed both for experienced devs and to give new devs a focus on _what_ to learn (thinking in systems) in the age of AI.

<p>Trace is a markdown-based declarative modeling approach to spec writing for LLM consumption.

It's designed both for experienced devs and to give new devs a focus on _what_ to learn (thinking in systems) in the age of AI.</p>