An autonomous AI coding loop designed to automate the completion of software development tasks. It reads requirements from a PRD file, YAML, or GitHub Issues, then delegates tasks to various AI coding assistants (Claude, Codex, OpenCode, Cursor, Qwen) to implement features, write tests, and commit changes until the project is complete. This tool is beneficial for developers seeking to accelerate development cycles by automating repetitive coding tasks and managing complex workflows.

How It Works

Ralphy operates by parsing tasks from specified sources (Markdown PRD, YAML, or GitHub Issues). Each task is assigned to an AI agent, which works within an isolated Git worktree and branch. The AI implements the requested feature, generates tests, and commits the changes. Ralphy orchestrates these agents, either sequentially or in parallel, managing their progress and merging completed work back to a base branch. It can also automatically generate pull requests for each task, with AI assistance for resolving merge conflicts.

Quick Start & Requirements

• Primary Install/Run: Clone the repository (git clone https://github.com/michaelshimeles/ralphy.git), navigate into the directory (cd ralphy), make the script executable (chmod +x ralphy.sh), and run it (e.g., ./ralphy.sh).
• Prerequisites: Requires one of the following AI assistant CLIs installed and configured: Claude Code, OpenCode, Codex, Cursor (with agent in PATH), or Qwen-Code. Also requires jq for JSON parsing.
• Optional Dependencies: yq (for YAML task files), gh (for GitHub Issues/PR creation), bc (for cost calculation).
• Task Sources: Supports Markdown files (default PRD.md), YAML files (tasks.yaml), and GitHub Issues (owner/repo).
• Links: Project repository: https://github.com/michaelshimeles/ralphy

Highlighted Details

• Supports multiple AI backends including Claude Code, Codex, OpenCode, Cursor, and Qwen-Code.
• Enables parallel execution of tasks using isolated Git worktrees and branches, with configurable maximum concurrency.
• Offers flexible task input formats: Markdown, YAML, and GitHub Issues with optional labeling.
• Automated Git branching per task, with options to create standard or draft pull requests.
• Includes AI-powered merge conflict resolution and cost tracking metrics for supported engines.

Maintenance & Community

No specific information regarding maintainers, community channels (e.g., Discord, Slack), sponsorships, or a public roadmap is provided in the README.

Licensing & Compatibility

• License: MIT License.
• Compatibility: The MIT license is permissive, generally allowing for commercial use and integration within closed-source projects without significant restrictions.

Limitations & Caveats

The tool's effectiveness is contingent on the correct installation and configuration of the chosen AI assistant CLIs. The quality of AI-generated code and test results directly impacts the success of the automated workflow. AI-driven merge conflict resolution may not always be successful, potentially requiring manual intervention. The project relies heavily on external AI models, and their inherent limitations or biases may be reflected in the output.