FA

fazer-ai/n8n-to-langgraph

开发工具

本页由 AI 翻译,技术细节请以原文为准。 查看原文

6 stars 质量 45 趋势 14

Claude Code plugin for converting n8n workflows to LangGraph TypeScript applications

概览

A Claude Code marketplace plugin that converts n8n workflows to LangGraph TypeScript applications. This plugin provides a structured pipeline for converting n8n workflow JSON files into fully functional LangGraph applications with . It handles the full lifecycle: analysis, architecture planning, implementation guidance, and post-implementation review. Every converted application includes a Langfuse helper module that traces all LLM calls, tool invocations, and graph executions — giving you a dashboard with latency, token usage, costs, and session replay. The integration is opt-in via environment variables: zero overhead when keys are not set. 3 specialized subagents + 1 orchestrator skill: Each agent has domain-specific skills preloaded via frontmatter — they don't inherit from the parent conversation. This guarantees the right knowledge is available regardless of how the agent is invoked. See the plugin marketplaces docs for more details. Alternatively, add to your ~/.

README

以下为原文 README 节选,完整内容请查看原文。

n8n-to-langgraph

A Claude Code marketplace plugin that converts n8n workflows to LangGraph TypeScript applications.

What it does

This plugin provides a structured pipeline for converting n8n workflow JSON files into fully functional LangGraph applications with Langfuse observability built in by default. It handles the full lifecycle: analysis, architecture planning, implementation guidance, and post-implementation review.

Every converted application includes a Langfuse helper module that traces all LLM calls, tool invocations, and graph executions — giving you a dashboard with latency, token usage, costs, and session replay. The integration is opt-in via environment variables: zero overhead when keys are not set.

Architecture

3 specialized subagents + 1 orchestrator skill:

/n8n-to-langgraph 
│
├─► workflow-analyzer (×2-3 in parallel)
│   Parses n8n JSON, extracts system prompts verbatim,
│   traces execution paths, maps integrations & credentials
│
├─► conversion-architect
│   Designs LangGraph StateGraph topology, state schemas,
│   tool specs, test plan, and phased implementation order
│
└─► conversion-reviewer (review mode)
    Compares implementation against originals for fidelity,
    test coverage, and code quality

Each agent has domain-specific skills preloaded via frontmatter — they don’t inherit from the parent conversation. This guarantees the right knowledge is available regardless of how the agent is invoked.

Plugin structure

.claude-plugin/
├── plugin.json              # Plugin metadata
└── marketplace.json         # Marketplace registration
agents/
├── workflow-analyzer.md     # n8n workflow analysis agent
├── conversion-architect.md  # LangGraph architecture design agent
└── conversion-reviewer.md   # Post-implementation review agent
skills/
└── n8n-to-langgraph/
    └── SKILL.md             # Orchestrator skill + domain knowledge

Installation

In Claude Code, run:

/plugin marketplace add fazer-ai/n8n-to-langgraph
/plugin install n8n-to-langgraph@n8n-to-langgraph

See the plugin marketplaces docs for more details.

Manual installation

Alternatively, add to your ~/.claude/settings.json:

{
  "enabledPlugins": {
    "n8n-to-langgraph@n8n-to-langgraph": true
  },
  "extraKnownMarketplaces": {
    "n8n-to-langgraph": {
      "source": {
        "source": "git",
        "url": "https://github.com/fazer-ai/n8n-to-langgraph.git"
      }
    }
  }
}

Then restart Claude Code.

Usage

Convert workflows

/n8n-to-langgraph workflows/

This runs two phases:

  1. Analysis — Parallel agents parse your n8n workflow JSONs, extract every system prompt verbatim, trace execution paths, and identify all integrations and credentials needed.
在 GitHub 查看完整 README

推荐工具

换一个关键词,或者移除筛选条件。

安装

npx skillfish add fazer-ai/n8n-to-langgraph
新手?先看 mcpskill.org 教程 →