Sign inSign up

looptroopai/looptroop

By looptroopai

•Updated 2 days ago

Local AI coding orchestrator: LLM-council planning, Ralph loops, isolated Git worktrees.

Image
Machine learning & AI
0

1.7K

looptroopai/looptroop repository overview

⁠LoopTroop

Local-first, transparent AI coding orchestration. Plan with multi-LLM councils, execute in isolated Git worktrees, and recover automatically with Ralph loops.

GitHub License: MIT Documentation


⁠Quickstart

Run LoopTroop locally with persistent storage:

docker run -d \
  --name looptroop \
  -p 57321:57321 \
  -v looptroop_data:/app/data \
  -v /var/run/docker.sock:/var/run/docker.sock \
  looptroopai/looptroop:latest

Open http://localhost:57321 in your browser.

⁠Docker Compose Example
version: '3.8'
services:
  looptroop:
    image: looptroopai/looptroop:latest
    container_name: looptroop
    ports:
      - "57321:57321"
    volumes:
      - looptroop_data:/app/data
      - /var/run/docker.sock:/var/run/docker.sock
    restart: unless-stopped

volumes:
  looptroop_data:

⁠Key Features

  • LLM Council Planning: Multiple models debate, vote on, and assemble execution plans.
  • Isolated Worktrees: Every task executes in dedicated Git worktrees without touching your working branch.
  • Ralph Loops: Automated test, lint, and error recovery before code is merged.
  • Transparent & Local: Full audit logs, local SQLite storage, and direct control over your workspace.

⁠Configuration & Ports

  • Port 57321: Main web UI and local server API.
  • Volume /app/data: Persistent storage for SQLite database state and project history. Always mount a named volume or host directory so data persists across container updates.

Tag summary

Content type

Image

Digest

sha256:755ba3d23…

Size

136.9 MB

Last updated

2 days ago

docker pull looptroopai/looptroop