LibreAI / Dev Tools / ralph-claude-code
// devtools · frankbria/ralph-claude-code

ralph-claude-code

An autonomous loop for Claude Code with intelligent exit detection, rate limits and a circuit breaker.

★ 9.7k stars⑂ 726 forksMITupdated 2026-10
▦ visual — this project's README ships no embeddable screenshot,
so here's a placeholder. The guide below is complete regardless.
// project facts

Project facts

ralph-claude-code is an open-source dev tools project by frankbria (frankbria/ralph-claude-code on GitHub). It has 9,671 GitHub stars and 726 forks, is released under the MIT license, and was last updated on GitHub on 8 October 2026.

GitHub stars
9,671
Forks
726
License
MIT
Last updated on GitHub
8 October 2026
Category
Dev Tools
Listing
Full install guide
// how it works

How it works

  1. Clarify your goal

    Define the project and task list Ralph should work through, and set sensible call limits in the project config before letting it run unattended.

  2. Get the project

    Clone frankbria/ralph-claude-code from GitHub (or use the install path in the run steps below) and skim its README for the parts relevant to your goal.

  3. Install and configure

    Clone the repo, run its install script, and make sure the Claude Code CLI and GitHub CLI are installed; then enable Ralph in a project with the setup wizard.

  4. Run your first task

    Start the loop in a project, watch live output with the live flag, and let the exit gate and circuit breaker decide when it is done or stuck.

// features

Why people love it

Dual-condition exit gate

Exit requires BOTH completion indicators AND an explicit EXIT_SIGNAL, so the loop neither quits early nor claims success falsely.

Rate limiting

An hourly limiter (100 calls/hour by default, configurable) with hourly reset keeps unattended runs inside a budget.

Circuit breaker

Advanced error detection trips a circuit breaker to prevent runaway loops when the agent is stuck repeating failures.

Response analyzer

Semantic response analysis with two-stage error filtering, plus JSON output format support with automatic fallback to text parsing.

Session continuity

A --resume flag preserves context across runs with configurable session expiration (24 hours by default), without session hijacking.

Operator tooling

Live streaming output, dry-run mode, desktop notifications, log rotation, ralph-stats metrics, and automatic git backup branches with rollback.

Advertisement
// run it

Run it yourself

  1. Clone the repository

    Get the source from the official GitHub repository for ralph-claude-code.

    git clone https://github.com/frankbria/ralph-claude-code.git
    cd ralph-claude-code
  2. Install Ralph

    Run the project's own installer after cloning, as the README documents.

    ./install.sh
  3. Enable and run in a project

    Use the enable wizard to set up a project, then start the autonomous loop.

    ralph-setup my-project  # via the ralph-enable wizard
    ralph --live  # see README for the exact loop command in your setup
// use cases

What you can build

Advertisement
// faq

Questions, answered

Who maintains ralph-claude-code?

frankbria, through the frankbria/ralph-claude-code repository.

Is ralph-claude-code free to use?

Yes — the project is open source under the MIT license as recorded for this listing. You can use, modify and distribute it under the MIT terms.

Does ralph-claude-code need a paid API key?

Yes — it drives the Claude Code CLI, so you need Claude Code set up with your own Anthropic access; Ralph itself is free and open source.

How does Ralph compare to just running Claude Code interactively?

Interactive Claude Code stops whenever you stop prompting. Ralph wraps it in a managed loop — exit detection, rate limiting, a circuit breaker, metrics and git backups — designed for unattended runs, at the cost of needing careful task setup.

How do I get started with ralph-claude-code?

Clone the repository (or use the documented installer), follow the install and configure steps on this page, and run the smallest real task first — start the loop in a project, watch live output with the live flag, and let the exit gate and circuit breaker decide when it is done or stuck.

// keep exploring

Similar tools

Dev Tools

claudex-loop

Claude Code skill: four-phase plan hardening (recon, interrogate, Codex adversarial review, cross-model build & inspection) — two AI models harden your plan …

★ 2.7k starsUnknown
View guide →
Dev Tools

flounder

Autonomous white-hat auditor — AI code review, bug-bounty research, exploit checks.

★ 516 starsAGPL-3.0
View guide →
Dev Tools

Mobile-Harness

Claude Code on Android — an AI mobile coding IDE with Linux commands.

★ 506 starsMIT
View guide →
Dev Tools

repowise

Repowise — codebase intelligence: health scores, auto-docs, git analytics and dead-code detection via MCP.

★ 7.2k starsAGPL-3.0
View guide →
Dev Tools

clawgod

Runtime patch for Claude Code, kept compatible with every version update.

★ 2.1k starsGPL-3.0
View guide →
Dev Tools

jarvis

A voice assistant for Claude Code — brainstorm, build and hear it report back.

★ 824 starsCustom
View guide →
// alternatives

Alternatives in Dev Tools

The most-starred other Dev Tools projects in the LibreAI catalogue — the realistic shortlist if ralph-claude-code is not the right fit.

Get ralph-claude-code

Free and open — grab the source, star the repo, and start creating.