# umputun/ralphex

Extended Ralph loop for autonomous AI-driven plan execution

Repository: https://github.com/umputun/ralphex
Canonical: https://ross.abutalabs.com/products/ralphex
Homepage: https://ralphex.com
Language: Go
License: MIT
License Family: permissive
Topics: automation, claude-code, codex, ralph-loop, ralph-wiggum, review
Last push: 2026-08-26T18:03:51+00:00

## Health v2 (maintenance only)
Score: 76/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 99, release rhythm 82, longevity 16
- inputs: {"age_days": 226, "days_push": 7, "days_rel": 43, "gap_med": 0, "n_releases_24m": 76}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1454, forks 119 (observed 2026-08-28T04:04:46.760527+00:00)

## What it is
ralphex is a standalone Go CLI tool that orchestrates coding agents like Claude Code and codex to autonomously execute structured implementation plans. Each task runs in a fresh agent session to avoid context degradation, followed by multi-phase AI code review and automatic git commits.

## Use cases
- run claude code autonomously on a multi-task implementation plan
- execute a coding plan overnight without babysitting the agent
- keep ai coding agent context fresh across long features
- automate multi-agent code review of ai-generated code
- auto-commit and track progress of ai-driven development tasks
- get notified when an autonomous coding run finishes or fails
- monitor autonomous agent execution from a web dashboard

## When to choose
- you have a structured plan with tasks and validation commands to execute via Claude Code or codex
- long interactive agent sessions degrade quality and you want fresh context per task
- you want automated multi-agent code review and commits after each task
- you prefer a single binary with zero setup and optional Docker isolation

## When to avoid
- you need an IDE plugin or GUI-driven coding assistant
- your workflow requires human approval at every step
- you don't use git or a compatible coding agent like Claude Code or codex
- you want the tool to write code itself rather than orchestrate existing agents

## Facets
- artifact type: cli-tool
- maturity: active
- function: agent-framework, workflow-automation, developer-tools, cli, code-review
- domain: developer-tools, large-language-models
- platform: cli, windows, go
- tags: ralph-loop, claude-code, codex, autonomous-coding, ai-orchestration, multi-agent-review, git-automation, ai-agents, automation, command-line, linux, macos, docker

## Member repositories
- umputun/ralphex (main) score 76

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:46.760527+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:35:43.050289+00:00, confidence not recorded.
  - readme: https://github.com/umputun/ralphex (fetched 2026-08-28T04:04:46.760527+00:00, sha 34a3f941b8af)
  - homepage: https://ralphex.com (fetched 2026-08-29T11:44:56.026282+00:00, sha 537e2bffbb28)
  - site_page: https://ralphex.com/docs (fetched 2026-08-29T11:44:56.029152+00:00, sha 9fc32b24aceb)
- Data as of 2026-08-30T08:39:29.467469+00:00.