What problem it solves

AI coding tools can lose context between tasks and struggle to carry development work forward consistently. Ralph addresses this by starting each iteration with a clean context while retaining progress through git history, progress.txt, and prd.json.

Key features

• Fresh-context autonomous iterations — Each iteration launches a new AI instance with clean context. Work continuity comes from committed changes and project state rather than conversational history.

  Source: README, ralph.sh

• PRD-driven task selection — The loop selects the highest-priority story that has not passed and updates prd.json as stories are completed.

  Source: README, prd.json.example

• Quality-gated execution — Ralph implements one story at a time, runs type checks and tests, and commits changes when those checks pass.

  Source: README

• Persistent project learnings — The loop appends learnings to progress.txt and updates relevant AGENTS.md files so later iterations and human developers can use discovered conventions and gotchas.

  Source: README, AGENTS.md

• Multiple AI tool integrations — Ralph supports Amp and Claude Code through separate prompt templates and a tool-selection option. It can be copied into a project, installed as global skills, or added as a Claude Code marketplace plugin.

  Source: README, prompt.md, CLAUDE.md

Quick start

Copied verbatim from the project's README.

# From your project root
mkdir -p scripts/ralph
cp /path/to/ralph/ralph.sh scripts/ralph/

# Copy the prompt template for your AI tool of choice:
cp /path/to/ralph/prompt.md scripts/ralph/prompt.md    # For Amp
# OR
cp /path/to/ralph/CLAUDE.md scripts/ralph/CLAUDE.md    # For Claude Code

chmod +x scripts/ralph/ralph.sh

# Using Amp (default)
./scripts/ralph/ralph.sh [max_iterations]

# Using Claude Code
./scripts/ralph/ralph.sh --tool claude [max_iterations]

Who it may suit

This may suit developers who manage AI-assisted implementation from a PRD and want each task to run in a fresh context with explicit verification and persisted learnings.

Topics to discuss

• Which project state should Ralph persist between iterations beyond git history, progress.txt, and prd.json?

• How should Ralph balance small, context-sized stories with dependencies that span multiple tasks?

• Which quality checks and browser verification practices work best for different kinds of projects?

Written by ReadmeX AI from the project's README, source and public GitHub data; it may contain mistakes.

Read the original README https://github.com/snarktank/ralph#readme

ReadmeX opened this community for a trending project; it is not run by the project. Maintainers can claim it from the community page.