convez.ai
SkillAutomaçãoAvançado
portpowered

You Agent Factory

portpowered/you-agent-factory

Orquestrador de agentes IA com dashboard e execução concorrente

Ver no GitHub
InicialPluginScore · de 0 a 100

Engine de orquestração de workflows para agentes IA que executa múltiplos agentes concorrentemente. Permite codificar processos em workflows, rotear trabalho entre estações de trabalho e executar loops de revisão automatizados com suporte a backends como OpenAI Codex e Claude.

Instalação

$npx convez-ai portpowered/you-agent-factory

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
16
Forks
1
Acompanhando
0
Saúde
50%
Último commit
há 1 mês
Versão
localai-backends-v1-3dc488f3dc9848f8e245954d28a3cd223aca1be42003e739d66ee3dd3b423598
Issues abertas
168
Contribuidores
1

Casos de uso

  • Executar múltiplos agentes em paralelo para tarefas independentes
  • Implementar loops de escrita-revisão automatizados com re-fila de trabalho
  • Agendar tarefas em git worktrees com triggers cron
  • Rotear e orquestrar workflows complexos entre workstations

Veredito

Ferramental robusto em Go para orquestração de agentes IA em escala, com dashboard visual e suporte a múltiplos backends. Indicado para times que precisam automatizar workflows complexos de engenharia de software e processamento de IA, mas requer configuração técnica avançada.

README

A tradução para o português deste README ainda está na fila.

you-agent-factory

CI Latest Release Go Version License: MIT

you-agent-factory is an AI agent factory for scheduling and orchestrating concurrent AI work—the you CLI and dashboard let you run many agents at once instead of babysitting each task manually.

you-agent-factory dashboard showing concurrent agent workstations, live work routing, and factory flow status

Why?

Leverage.

With you-agent-factory, you codify your process into a workflow with different AGENTS.md files and run them as wrappers around OpenAI Codex or other agent backends.

For example:

  • dispatch many agents to run independently in separate worktrees
  • have one agent loop through a series of tasks, then route output to a reviewer that re-queues failed work
  • submit plans in dependency order
  • use cron triggers to autonomously inspect git tasks and drive write/review cycles

Installation

Prerequisites

  • Codex CLI (default agent backend for the starter factory): npm i -g @openai/codex
  • A project directory where you want the local factory/ scaffold to live

Install the you CLI

macOS / Linux:

curl -fsSL https://github.com/portpowered/you-agent-factory/releases/latest/download/install.sh | sh

Windows (PowerShell):

irm https://github.com/portpowered/you-agent-factory/releases/latest/download/install.ps1 | iex

For custom install locations or pinned versions, see the install script.

Quick start

The default path uses the Codex-backed starter scaffold:

  1. cd your-project-directory
  2. Run you — bootstraps ./factory, starts the runtime, and prints the dashboard URL (usually http://localhost:7437/dashboard/ui)
  3. Submit a task from the dashboard (for example, “write a report on my codebase to TEST.md”) and wait for completion

For factory authoring, CLI topics, and advanced setup, see Authoring factories and you docs.

Alternate executor: Claude

To scaffold a factory with Claude as the starter worker instead of Codex:

you init --executor claude --dir my-factory
you docs workstation

Features

you-agent-factory is a factory runtime: you define how work moves between workstations, and the you CLI plus dashboard schedule concurrent agent runs against that flow.

  • Concurrent agent execution — Run many agents at once across workstations; the dashboard shows live routing, session status, and factory flow state.
  • Workflow customization — Model processes as config (factory.json, workstation routes, AGENTS.md) instead of a fixed pipeline; adapt write/review loops, cron triggers, git worktrees, or other patterns to your repo.
  • Review loops — Route completed work to reviewer workstations and re-queue failed items; shipped factories such as Ralph and writer-reviewer demonstrate iterative plan/code/review cycles.
  • Batch submission — Submit single items from the CLI (you submit) or drive larger inputs through batch work types and dashboard submission.
  • Example factories — Load starter and advanced factories from examples/factories/ in the dashboard, or scaffold your own with you init.

Deeper product documentation:

Comparison

How you-agent-factory fits next to nearby agent and workflow orchestrators. Dimensions focus on execution model, workflow flexibility, agent-harness support, and operational weight—not “best tool” claims. For a longer maintainer write-up, see Comparing orchestration systems.

SystemExecution modelWorkflow shapeAgent harnessDurability / ops weightReference
you-agent-factorySelf-hosted you runtime and dashboard route work through factory workstationsCustom in-repo flow (factory.json, AGENTS.md, routes) without a fixed pipelineCodex, Claude, and shell workers wired through factory configLightweight local runtime; no built-in durable workflow engineArchitecture · Comparing systems
GastownMayor-led multi-agent workspace with git-backed hooks and worktreesOpinionated mayor/beads/convoy coordination around gitHooks inject context into Claude Code, Copilot, Codex, and peersGit/worktree persistence; heavier git + beads/dolt stackGastown
SymphonyLong-running orchestrator polls issue trackers and runs per-issue workspacesPolicy in-repo (WORKFLOW.md); spec-driven daemon workflowCodex app-server sessions in isolated workspacesElixir daemon with supervision/retries; tracker-centricSymphony
FactoryDroid Missions orchestrator with Mission Control for multi-day projectsMilestone/feature decomposition with validation contractsDroid workers with MCP, skills, hooks, and custom droidsProductized orchestration with milestone validation loopsFactory Missions
8090 Software FactoryHosted SDLC control plane (requirements → blueprints → work orders)Upstream planning modules feed agents through MCP-connected work ordersExternal agents (Cursor, Claude Code, etc.) via MCPCloud platform with knowledge graph and audit trail8090 docs
Claude workflow pluginsIn-IDE Claude Code skills/commands drive spec → plan → execute loopsPlugin-defined task graphs, parallel dispatch, and worktreesNative Claude Code subagents and skillsSession/task files on disk; no separate orchestration serversighup/claude-workflow
Other orchestrators (Temporal, n8n, DBOS)General-purpose durable or RPA workflow enginesDAG/state-machine or node graphs; often code- or node-config drivenAgent harnesses typically custom-builtStrong durability/transactions; heavier for ad-hoc agent loopsComparing systems

References

Maintainers: edit packaged reference docs under docs/reference/; run make docs-reference-smoke before shipping doc changes and make readme-check before changing README structure or linked assets.

License

This repository is released under the MIT License.

The README hero image (docs/internal/resources/dashboard.png) and the animated demo (docs/internal/resources/dashboard.gif) are screenshots maintained in this repository and depict the you-agent-factory dashboard UI.

Ver o original no GitHub

Bom para

Relacionados