Skip to main content
Glama

McFlow

path-reference.md4.61 kB
# McFlow Path Reference Guide ## ⚠️ DEFINITIVE SOURCE: See [ARCHITECTURE.md](/ARCHITECTURE.md) This is a quick reference. For the complete, authoritative path structure, see [ARCHITECTURE.md](/ARCHITECTURE.md). ## Key Points: - `projectPath` = Project root directory - `workflowsPath` = `{projectPath}/workflows/` - All McFlow paths are **relative to workflowsPath** ### Workflow Files **Internal location:** `{workflowsPath}/flows/` **Read a workflow:** ```bash mcflow read --path "flows/project-workflow-name.json" ``` **Examples:** ```bash ✅ mcflow read --path "flows/newspop-news-collection.json" # Correct - relative to workflowsPath ✅ mcflow read --path "flows/newspop-image-analysis.json" # Correct - relative to workflowsPath ❌ mcflow read --path "news-collection.json" # Missing flows/ subdirectory ❌ mcflow read --path "workflows/flows/news-collection.json" # Don't include workflows/ ❌ mcflow read --path "newspop-news-collection" # Missing flows/ and .json ``` ### Extracted Node Files **Code files:** `workflows/nodes/code/` ```bash workflows/nodes/code/workflow-name-node-purpose.js workflows/nodes/code/newspop-news-analysis-prepare-items.js ``` **Prompt files:** `workflows/nodes/prompts/` ```bash workflows/nodes/prompts/workflow-name-prompt-name.md workflows/nodes/prompts/newspop-analysis-batch.md ``` **JSON files:** `workflows/nodes/json/` ```bash workflows/nodes/json/sample-data.json workflows/nodes/json/api-config.json ``` **SQL files:** `workflows/nodes/sql/` ```bash workflows/nodes/sql/query-name.sql ``` ### Complete Directory Structure ``` project-root/ └── workflows/ # Main workflows directory ├── flows/ # Workflow JSON files │ ├── project-workflow-1.json │ ├── project-workflow-2.json │ └── project-workflow-3.json ├── nodes/ # Extracted node content │ ├── code/ # JavaScript/TypeScript files │ │ ├── workflow1-node1.js │ │ └── workflow2-node1.js │ ├── prompts/ # AI prompt files (Markdown) │ │ ├── analysis-prompt.md │ │ └── summary-prompt.md │ ├── json/ # JSON configurations and mock data │ │ ├── sample-articles.json │ │ └── api-config.json │ ├── sql/ # SQL queries │ │ └── fetch-users.sql │ └── templates/ # HTML/text templates │ └── email-template.html ├── dist/ # Compiled workflows (auto-generated) ├── .env # Environment variables (never commit!) └── README.md # Workflows documentation ``` ## Common Path Errors and Fixes ### Error: "ENOENT: no such file or directory" **Cause:** Using incomplete path ```bash ❌ mcflow read --path "image_analysis.json" ``` **Fix:** Use complete path ```bash ✅ mcflow read --path "workflows/flows/newspop-image-analysis.json" ``` ### Error: "Failed to read workflow" **Cause:** Missing `workflows/` prefix ```bash ❌ mcflow read --path "flows/workflow.json" ``` **Fix:** Include full path from project root ```bash ✅ mcflow read --path "workflows/flows/workflow.json" ``` ### Error: "Cannot find module" **Cause:** Wrong node file path ```bash ❌ const code = require('./node-code.js'); ``` **Fix:** Use McFlow's extraction system ```bash ✅ // File will be at: workflows/nodes/code/workflow-node.js // McFlow handles the injection automatically ``` ## Rules for AI Agents 1. **ALWAYS use full paths** starting with `workflows/` 2. **NEVER assume current directory** - always specify from project root 3. **CHECK path exists** before operations 4. **USE the list command** to see available workflows before reading 5. **FOLLOW the structure** - don't create files outside the standard locations ## Quick Commands **List all workflows:** ```bash mcflow list ``` **Read specific workflow:** ```bash mcflow read --path "workflows/flows/[project]-[workflow].json" ``` **Extract code from workflow:** ```bash mcflow extract_code --workflow "[project]-[workflow]" ``` **Deploy workflow:** ```bash mcflow deploy --path "workflows/flows/[project]-[workflow].json" ``` --- **Remember:** When in doubt, use `mcflow list` first to see the correct workflow names and paths!

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/mckinleymedia/mcflow-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server