McFlow
# McFlow
MCP server for managing n8n workflows with external file support.
## Installation
```bash
git clone https://github.com/mckinleymedia/mcflow-mcp.git
cd mcflow-mcp
npm install
npm run build
```
## Setup
Add to Claude Code config (`~/.claude.json`):
```json
{
"mcpServers": {
"mcflow": {
"type": "stdio",
"command": "node",
"args": [
"<ABSOLUTE_PATH_TO_REPO>/mcflow-mcp/dist/index.js"
],
"env": {}
}
}
}
```
## Features
- Extract code/prompts/SQL from nodes into editable files
- Deploy workflows with automatic code compilation
- Generate workflow templates
- Manage n8n instance (auto-installs if needed)
- **NEW**: Generic workflow tracking system with checkpoints and error handling
- **NEW**: Generate Next.js apps for workflow data management
## Docs
- [Getting Started](docs/overview.md)
- [AI Instructions](docs/ai/instructions.md)
- [Workflow Tracking](docs/workflow-tracking.md)
- [App Generation](docs/app-generation.md)
## License
MITTDQS
Scored across 27 tools
Most tools have distinct purposes, but some overlap exists. For example, 'list' and 'deployed' both list workflows but in different contexts (project vs n8n instance), which could cause confusion. Similarly, 'analyze' and 'validate' both involve workflow inspection but with different focuses, potentially leading to misselection.
The naming is mostly consistent with a verb-based pattern (e.g., 'activate', 'add_node', 'create', 'deploy'), but there are minor deviations like 'deployed' (past tense verb) and 'info' (noun). Overall, the naming is readable and follows a clear convention, with only a few inconsistencies.
With 27 tools, the count is too high for a focused workflow management server. This many tools suggests feature bloat and can overwhelm agents, making it difficult to navigate and select the right tool efficiently. A more streamlined set of 10-15 tools would be more appropriate.
The tool set provides comprehensive coverage for n8n workflow management, including creation, deployment, execution, analysis, and maintenance. It covers the full lifecycle from 'create' to 'deploy' and 'execute', with additional tools for code extraction, tracking, and validation, leaving no obvious gaps.