Ralph Wiggum, the viral agentic coding loop.Now simplified for everyone.
The open source approach that went viral for autonomous AI delivery, tuned for real-world teams. Community-led, MIT-licensed, and built to stay simple to adopt.
AI coding agents are powerful.But they need structure.
Modern AI agents can write entire applications. Without clear specifications, they wander, over-engineer, and lose focus on what actually matters.
Ralph Wiggum + SpecKitThe best of both worlds.
We combined the relentless iteration of Ralph Wiggum with a simplified version of the SpecKit methodology from GitHub.
How it works
A simple loop that ships.
The same process across Claude Code, Codex, Cursor, and more.
Create a feature specification using natural language. The AI generates professional acceptance criteria.
/speckit.specify Add user authentication with OAuth
The AI works autonomously until all criteria pass. It commits, pushes, deploys, tests, and iterates.
/speckit.implement
When everything passes, you get production-ready code that ships with confidence.
<promise>DONE</promise>
Why Ralph Wiggum
Built for real-world delivery.
Iterative self-correction
The AI does not stop at the first attempt. It tests, finds issues, fixes them, and repeats until everything actually works.
Spec-driven development
Professional-grade specifications ensure the AI knows exactly what to build. No scope creep or forgotten requirements.
Works everywhere
Claude Code, OpenAI Codex, Cursor, and more. One approach, any platform.
Instant setup
Point your AI to the repo. That is it. You are ready to build.
Open source community
Built in public, for the public.
Ralph Wiggum lives on GitHub, shaped by the open source community and shared with a non-profit spirit. Fork it, remix it, and ship better specs with the world.
Transparent, forkable, and built to empower contributors.
Ideas evolve in the open and stay friendly to every builder.
Share it freely, use it commercially, and keep it moving.
Tree of Ralphs: nested autonomous loops
For complex projects with multiple specs, Ralph Wiggum uses a tree structure. Each spec gets its own loop, and the master loop orchestrates everything until the entire project is done.
Master Ralph (All Specs) | +-- Ralph 001 (Spec 001) -> <DONE> +-- Ralph 002 (Spec 002) -> <DONE> +-- Ralph 003 (Spec 003) -> <DONE> | +-- <ALL_DONE>
Getting started
Start building in 60 seconds.
Paste either prompt into your AI agent and you are ready to go.
New project
Use this when starting from scratch.
Existing project
Use this to add Ralph Wiggum to an existing repo.
How we improved SpecKit
SpecKit is excellent for spec-driven development, but it was designed before modern AI agents became so capable. Here is what we simplified.
| SpecKit Step | Our Approach | Why |
|---|---|---|
| /speckit.constitution | Keep | Essential project principles |
| /speckit.specify | Keep (enhanced) | Specs with Completion Signals |
| /speckit.plan | Skip | AI agents plan dynamically |
| /speckit.tasks | Skip | AI agents break down work automatically |
| /speckit.implement | Replace with Ralph loop | Iterative until acceptance criteria pass |
| specify-cli | Skip | Prompts are enough |
Standing on the shoulders of giants
Ralph Wiggum builds on incredible work by the community. This is our interpretation of those ideas, and what we believe is a more approachable update for modern agentic coding.