Compare commits

..

1 Commits

Author SHA1 Message Date
Claude
fe718f62c5 fix: replace YAML-conflicting markdown separators in release workflow
Changed Markdown horizontal rules from "---" to "___" in heredoc content
to prevent YAML parser conflicts. The "---" string at the start of a line
is a YAML document separator which causes parsing issues even inside bash
heredocs within GitHub Actions workflows.

Root cause:
- Lines 151 and 226 contained "---" Markdown horizontal rules
- YAML parsers interpret "---" as document separators
- This prevented workflow jobs from executing (no jobs scheduled)
- Changed to "___" which creates identical Markdown output

Impact:
- Fixes workflow run 18804889309 where no jobs executed
- Fixes workflow run 18804651253 with same issue
- Both "___" and "---" render as horizontal rules in Markdown
- Workflow should now properly detect version changes and execute all jobs

Related: Commit 79ef853 which removed emoji from heredoc

Concieved by Romuald Członkowski - www.aiadvisors.pl/en

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
2025-10-25 17:18:20 +00:00
19 changed files with 1791 additions and 905 deletions

View File

@@ -142,13 +142,19 @@ jobs:
if [ -z "$PREVIOUS_TAG" ]; then
echo " No previous tag found, this might be the first release"
# Generate initial release notes using script
if NOTES=$(node scripts/generate-initial-release-notes.js "$CURRENT_VERSION" 2>/dev/null); then
echo "✅ Successfully generated initial release notes for version $CURRENT_VERSION"
else
echo "⚠️ Could not generate initial release notes for version $CURRENT_VERSION"
NOTES="Initial release v$CURRENT_VERSION"
fi
# Get all commits up to current commit - use heredoc for multiline
NOTES=$(cat <<EOF
### Initial Release
This is the initial release of n8n-mcp v$CURRENT_VERSION.
___
**Release Statistics:**
- Commit count: $(git rev-list --count HEAD)
- First release setup
EOF
)
echo "has-notes=true" >> $GITHUB_OUTPUT
@@ -214,11 +220,11 @@ jobs:
# Create release body
cat > release_body.md << 'EOF'
# Release v${{ needs.detect-version-change.outputs.new-version }}
${{ needs.generate-release-notes.outputs.release-notes }}
---
___
## Installation
### NPM Package

View File

@@ -7,155 +7,6 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
## [Unreleased]
## [2.22.8] - 2025-11-03
### 🔄 Dependencies Update
**n8n Platform Update to 1.118.1**
Updated n8n and all related dependencies to the latest versions:
- **n8n**: 1.117.2 → 1.118.1
- **n8n-core**: 1.116.0 → 1.117.0
- **n8n-workflow**: 1.114.0 → 1.115.0
- **@n8n/n8n-nodes-langchain**: 1.116.2 → 1.117.0
### 📊 Database Changes
- Rebuilt node database with **542 nodes**
- 439 nodes from n8n-nodes-base
- 103 nodes from @n8n/n8n-nodes-langchain
- All node metadata synchronized with latest n8n release
### ✅ Verification
- Database rebuild completed successfully
- All node types validated
- Documentation mappings updated
Conceived by Romuald Członkowski - https://www.aiadvisors.pl/en
## [2.22.7] - 2025-10-26
### 📝 Documentation Fixes
**Issue #292: Corrected Array Property Removal Documentation in n8n_update_partial_workflow**
Fixed critical documentation error in property removal patterns that could have led users to write non-functional code.
#### Problem
The documentation incorrectly showed using array index notation `[0]` for removing array elements:
```javascript
// INCORRECT (doesn't work as shown)
updates: { "parameters.headers[0]": undefined }
```
**Root Cause**: The `setNestedProperty` implementation doesn't parse array index notation like `[0]`. It treats `headers[0]` as a literal object key, not an array index.
**Impact**: Users following the documentation would write code that doesn't behave as expected. The property `headers[0]` would be treated as an object key, not an array element reference.
#### Fixed
**Three documentation corrections in `src/mcp/tool-docs/workflow_management/n8n-update-partial-workflow.ts`:**
1. **Nested Property Removal Section** (lines 236-244):
- Changed comment from `// Remove array element` to `// Remove entire array property`
- Changed code from `"parameters.headers[0]": undefined` to `"parameters.headers": undefined`
2. **Example rm5** (line 340):
- Changed comment from `// Remove array element` to `// Remove entire array property`
- Changed code from `"parameters.headers[0]": undefined` to `"parameters.headers": undefined`
3. **Pitfalls Section** (line 405):
- OLD: `'Array element removal with undefined removes the element at that index, which may shift subsequent indices'`
- NEW: `'Array index notation (e.g., "parameters.headers[0]") is not supported - remove the entire array property instead'`
#### Correct Usage
**To remove an array property:**
```javascript
// Correct: Remove entire array
n8n_update_partial_workflow({
id: "wf_012",
operations: [{
type: "updateNode",
nodeName: "HTTP Request",
updates: { "parameters.headers": undefined } // Remove entire headers array
}]
});
```
**NOT:**
```javascript
// Incorrect: Array index notation doesn't work
updates: { "parameters.headers[0]": undefined } // Treated as object key "headers[0]"
```
#### Impact
- **Prevents User Confusion**: Clear documentation on what works vs. what doesn't
- **Accurate Examples**: All examples now show correct, working patterns
- **Better Error Prevention**: Pitfall warning helps users avoid this mistake
- **No Code Changes**: This is purely a documentation fix - no implementation changes needed
#### Testing
- ✅ Documentation reviewed by code-reviewer agent
- ✅ Tested by n8n-mcp-tester agent
- ✅ All examples verified against actual implementation behavior
- ✅ Pitfall accurately describes technical limitation
#### Files Changed
**Documentation (1 file)**:
- `src/mcp/tool-docs/workflow_management/n8n-update-partial-workflow.ts` - Corrected 3 instances of array property removal documentation
**Configuration (2 files)**:
- `package.json` - Version bump to 2.22.7
- `package.runtime.json` - Version bump to 2.22.7
#### Related
- **Issue**: #292 - Missing documentation on how to remove node properties using `updateNode`
- **PR**: #375 - Resolve GitHub Issue 292 in n8n-mcp
- **Code Review**: Identified critical array index notation documentation error
- **Root Cause**: Implementation doesn't parse array bracket notation `[N]`
Conceived by Romuald Członkowski - [www.aiadvisors.pl/en](https://www.aiadvisors.pl/en)
## [2.22.6] - 2025-10-25
### 🐛 Bug Fixes
**Issue #228: Fix Docker Port Configuration Mismatch**
Fixed critical Docker configuration bug where custom PORT environment variable values were not properly mapped to container ports, causing connection failures in Docker deployments.
#### Problem
- **docker-compose.yml**: Port mapping `"${PORT:-3000}:3000"` hardcoded container port to 3000
- **docker-compose.yml**: Health check hardcoded to port 3000
- **Dockerfile**: Health check hardcoded to port 3000
- Impact: When PORT≠3000 (e.g., PORT=8080), Docker mapped host port to wrong container port
#### Solution
- **docker-compose.yml line 44**: Changed port mapping to `"${PORT:-3000}:${PORT:-3000}"`
- **docker-compose.yml line 56**: Updated health check to use dynamic port `$${PORT:-3000}`
- **Dockerfile line 93**: Updated HEALTHCHECK to use dynamic port `${PORT:-3000}`
- **Dockerfile line 85**: Added clarifying comment about PORT configurability
#### Testing
- Verified with default PORT (3000)
- Verified with custom PORT (8080)
- Health checks work correctly in both scenarios
#### Related Issues
- Fixes #228 (Docker Compose port error)
- Likely fixes #109 (Configuration ignored in HTTP mode)
- Likely fixes #84 (Can't access container)
Conceived by Romuald Członkowski - [www.aiadvisors.pl/en](https://www.aiadvisors.pl/en)
## [2.22.3] - 2025-10-25
### 🔧 Code Quality Improvements

View File

@@ -82,7 +82,7 @@ ENV IS_DOCKER=true
# To opt-out, uncomment the following line:
# ENV N8N_MCP_TELEMETRY_DISABLED=true
# Expose HTTP port (default 3000, configurable via PORT environment variable at runtime)
# Expose HTTP port
EXPOSE 3000
# Set stop signal to SIGTERM (default, but explicit is better)
@@ -90,7 +90,7 @@ STOPSIGNAL SIGTERM
# Health check
HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 \
CMD sh -c 'curl -f http://127.0.0.1:${PORT:-3000}/health || exit 1'
CMD curl -f http://127.0.0.1:3000/health || exit 1
# Optimized entrypoint
ENTRYPOINT ["/usr/local/bin/docker-entrypoint.sh"]

View File

@@ -5,23 +5,23 @@
[![npm version](https://img.shields.io/npm/v/n8n-mcp.svg)](https://www.npmjs.com/package/n8n-mcp)
[![codecov](https://codecov.io/gh/czlonkowski/n8n-mcp/graph/badge.svg?token=YOUR_TOKEN)](https://codecov.io/gh/czlonkowski/n8n-mcp)
[![Tests](https://img.shields.io/badge/tests-3336%20passing-brightgreen.svg)](https://github.com/czlonkowski/n8n-mcp/actions)
[![n8n version](https://img.shields.io/badge/n8n-1.118.1-orange.svg)](https://github.com/n8n-io/n8n)
[![n8n version](https://img.shields.io/badge/n8n-^1.116.2-orange.svg)](https://github.com/n8n-io/n8n)
[![Docker](https://img.shields.io/badge/docker-ghcr.io%2Fczlonkowski%2Fn8n--mcp-green.svg)](https://github.com/czlonkowski/n8n-mcp/pkgs/container/n8n-mcp)
[![Deploy on Railway](https://railway.com/button.svg)](https://railway.com/deploy/n8n-mcp?referralCode=n8n-mcp)
A Model Context Protocol (MCP) server that provides AI assistants with comprehensive access to n8n node documentation, properties, and operations. Deploy in minutes to give Claude and other AI assistants deep knowledge about n8n's 541 workflow automation nodes.
A Model Context Protocol (MCP) server that provides AI assistants with comprehensive access to n8n node documentation, properties, and operations. Deploy in minutes to give Claude and other AI assistants deep knowledge about n8n's 525+ workflow automation nodes.
## Overview
n8n-MCP serves as a bridge between n8n's workflow automation platform and AI models, enabling them to understand and work with n8n nodes effectively. It provides structured access to:
- 📚 **541 n8n nodes** from both n8n-nodes-base and @n8n/n8n-nodes-langchain
- 📚 **536 n8n nodes** from both n8n-nodes-base and @n8n/n8n-nodes-langchain
- 🔧 **Node properties** - 99% coverage with detailed schemas
-**Node operations** - 63.6% coverage of available actions
- 📄 **Documentation** - 87% coverage from official n8n docs (including AI nodes)
- 🤖 **AI tools** - 271 AI-capable nodes detected with full documentation
- 📄 **Documentation** - 90% coverage from official n8n docs (including AI nodes)
- 🤖 **AI tools** - 263 AI-capable nodes detected with full documentation
- 💡 **Real-world examples** - 2,646 pre-extracted configurations from popular templates
- 🎯 **Template library** - 2,709 workflow templates with 100% metadata coverage
- 🎯 **Template library** - 2,500+ workflow templates with smart filtering
## ⚠️ Important Safety Warning
@@ -51,8 +51,6 @@ npx n8n-mcp
Add to Claude Desktop config:
> ⚠️ **Important**: The `MCP_MODE: "stdio"` environment variable is **required** for Claude Desktop. Without it, you will see JSON parsing errors like `"Unexpected token..."` in the UI. This variable ensures that only JSON-RPC messages are sent to stdout, preventing debug logs from interfering with the protocol.
**Basic configuration (documentation tools only):**
```json
{
@@ -533,7 +531,7 @@ When operations are independent, execute them in parallel for maximum performanc
❌ BAD: Sequential tool calls (await each one before the next)
### 3. Templates First
ALWAYS check templates before building from scratch (2,709 available).
ALWAYS check templates before building from scratch (2,500+ available).
### 4. Multi-Level Validation
Use validate_node_minimal → validate_node_operation → validate_workflow pattern.
@@ -842,7 +840,7 @@ n8n_update_partial_workflow({
### Core Behavior
1. **Silent execution** - No commentary between tools
2. **Parallel by default** - Execute independent operations simultaneously
3. **Templates first** - Always check before building (2,709 available)
3. **Templates first** - Always check before building (2,500+ available)
4. **Multi-level validation** - Quick check → Full validation → Workflow validation
5. **Never trust defaults** - Explicitly configure ALL parameters
@@ -945,7 +943,7 @@ Once connected, Claude can use these powerful tools:
- **`get_node_as_tool_info`** - Get guidance on using any node as an AI tool
### Template Tools
- **`list_templates`** - Browse all templates with descriptions and optional metadata (2,709 templates)
- **`list_templates`** - Browse all templates with descriptions and optional metadata (2,500+ templates)
- **`search_templates`** - Text search across template names and descriptions
- **`search_templates_by_metadata`** - Advanced filtering by complexity, setup time, services, audience
- **`list_node_templates`** - Find templates using specific nodes
@@ -1100,17 +1098,17 @@ npm run dev:http # HTTP dev mode
## 📊 Metrics & Coverage
Current database coverage (n8n v1.117.2):
Current database coverage (n8n v1.113.3):
- ✅ **541/541** nodes loaded (100%)
- ✅ **541** nodes with properties (100%)
- ✅ **470** nodes with documentation (87%)
- ✅ **271** AI-capable tools detected
- ✅ **536/536** nodes loaded (100%)
- ✅ **528** nodes with properties (98.7%)
- ✅ **470** nodes with documentation (88%)
- ✅ **267** AI-capable tools detected
- ✅ **2,646** pre-extracted template configurations
- ✅ **2,709** workflow templates available (100% metadata coverage)
- ✅ **2,500+** workflow templates available
- ✅ **AI Agent & LangChain nodes** fully documented
- ⚡ **Average response time**: ~12ms
- 💾 **Database size**: ~68MB (includes templates with metadata)
- 💾 **Database size**: ~15MB (optimized)
## 🔄 Recent Updates

Binary file not shown.

View File

@@ -20,19 +20,19 @@ services:
image: n8n-mcp:latest
container_name: n8n-mcp
ports:
- "${PORT:-3000}:${PORT:-3000}"
- "3000:3000"
environment:
- MCP_MODE=${MCP_MODE:-http}
- AUTH_TOKEN=${AUTH_TOKEN}
- NODE_ENV=${NODE_ENV:-production}
- LOG_LEVEL=${LOG_LEVEL:-info}
- PORT=${PORT:-3000}
- PORT=3000
volumes:
# Mount data directory for persistence
- ./data:/app/data
restart: unless-stopped
healthcheck:
test: ["CMD", "sh", "-c", "curl -f http://localhost:$${PORT:-3000}/health"]
test: ["CMD", "curl", "-f", "http://localhost:3000/health"]
interval: 30s
timeout: 10s
retries: 3

View File

@@ -37,12 +37,11 @@ services:
container_name: n8n-mcp
restart: unless-stopped
ports:
- "${MCP_PORT:-3000}:${MCP_PORT:-3000}"
- "${MCP_PORT:-3000}:3000"
environment:
- NODE_ENV=production
- N8N_MODE=true
- MCP_MODE=http
- PORT=${MCP_PORT:-3000}
- N8N_API_URL=http://n8n:5678
- N8N_API_KEY=${N8N_API_KEY}
- MCP_AUTH_TOKEN=${MCP_AUTH_TOKEN}
@@ -57,7 +56,7 @@ services:
n8n:
condition: service_healthy
healthcheck:
test: ["CMD", "sh", "-c", "curl -f http://localhost:$${MCP_PORT:-3000}/health"]
test: ["CMD", "curl", "-f", "http://localhost:3000/health"]
interval: 30s
timeout: 10s
retries: 3

View File

@@ -41,7 +41,7 @@ services:
# Port mapping
ports:
- "${PORT:-3000}:${PORT:-3000}"
- "${PORT:-3000}:3000"
# Resource limits
deploy:
@@ -53,7 +53,7 @@ services:
# Health check
healthcheck:
test: ["CMD", "sh", "-c", "curl -f http://127.0.0.1:$${PORT:-3000}/health"]
test: ["CMD", "curl", "-f", "http://127.0.0.1:3000/health"]
interval: 30s
timeout: 10s
retries: 3

View File

@@ -4,9 +4,7 @@ Connect n8n-MCP to Claude Code CLI for enhanced n8n workflow development from th
## Quick Setup via CLI
### Basic configuration (documentation tools only)
**For Linux, macOS, or Windows (WSL/Git Bash):**
### Basic configuration (documentation tools only):
```bash
claude mcp add n8n-mcp \
-e MCP_MODE=stdio \
@@ -15,21 +13,9 @@ claude mcp add n8n-mcp \
-- npx n8n-mcp
```
**For native Windows PowerShell:**
```powershell
# Note: The backtick ` is PowerShell's line continuation character.
claude mcp add n8n-mcp `
'-e MCP_MODE=stdio' `
'-e LOG_LEVEL=error' `
'-e DISABLE_CONSOLE_OUTPUT=true' `
-- npx n8n-mcp
```
![Adding n8n-MCP server in Claude Code](./img/cc_command.png)
### Full configuration (with n8n management tools)
**For Linux, macOS, or Windows (WSL/Git Bash):**
### Full configuration (with n8n management tools):
```bash
claude mcp add n8n-mcp \
-e MCP_MODE=stdio \
@@ -40,18 +26,6 @@ claude mcp add n8n-mcp \
-- npx n8n-mcp
```
**For native Windows PowerShell:**
```powershell
# Note: The backtick ` is PowerShell's line continuation character.
claude mcp add n8n-mcp `
'-e MCP_MODE=stdio' `
'-e LOG_LEVEL=error' `
'-e DISABLE_CONSOLE_OUTPUT=true' `
'-e N8N_API_URL=https://your-n8n-instance.com' `
'-e N8N_API_KEY=your-api-key' `
-- npx n8n-mcp
```
Make sure to replace `https://your-n8n-instance.com` with your actual n8n URL and `your-api-key` with your n8n API key.
## Alternative Setup Methods
@@ -159,11 +133,9 @@ For optimal results, create a `CLAUDE.md` file in your project root with the ins
## Tips
- If you're running n8n locally, use `http://localhost:5678` as the `N8N_API_URL`.
- The n8n API credentials are optional. Without them, you'll only have access to documentation and validation tools. With credentials, you get full workflow management capabilities.
- **Scope Management:**
- By default, `claude mcp add` uses `--scope local` (also called "user scope"), which saves the configuration to your global user settings and keeps API keys private.
- To share the configuration with your team, use `--scope project`. This saves the configuration to a `.mcp.json` file in your project's root directory.
- **Switching Scope:** The cleanest method is to `remove` the server and then `add` it back with the desired scope flag (e.g., `claude mcp remove n8n-mcp` followed by `claude mcp add n8n-mcp --scope project`).
- **Manual Switching (Advanced):** You can manually edit your `.claude.json` file (e.g., `C:\Users\YourName\.claude.json`). To switch, cut the `"n8n-mcp": { ... }` block from the top-level `"mcpServers"` object (user scope) and paste it into the nested `"mcpServers"` object under your project's path key (project scope), or vice versa. **Important:** You may need to restart Claude Code for manual changes to take effect.
- Claude Code will automatically start the MCP server when you begin a conversation.
- If you're running n8n locally, use `http://localhost:5678` as the N8N_API_URL
- The n8n API credentials are optional - without them, you'll have documentation and validation tools only
- With API credentials, you'll get full workflow management capabilities
- Use `--scope local` (default) to keep your API credentials private
- Use `--scope project` to share configuration with your team (put credentials in environment variables)
- Claude Code will automatically start the MCP server when you begin a conversation

View File

@@ -59,10 +59,10 @@ docker compose up -d
- n8n-mcp-data:/app/data
ports:
- "${PORT:-3000}:${PORT:-3000}"
- "${PORT:-3000}:3000"
healthcheck:
test: ["CMD", "sh", "-c", "curl -f http://127.0.0.1:$${PORT:-3000}/health"]
test: ["CMD", "curl", "-f", "http://127.0.0.1:3000/health"]
interval: 30s
timeout: 10s
retries: 3

1925
package-lock.json generated

File diff suppressed because it is too large Load Diff

View File

@@ -1,6 +1,6 @@
{
"name": "n8n-mcp",
"version": "2.22.8",
"version": "2.22.5",
"description": "Integration between n8n workflow automation and Model Context Protocol (MCP)",
"main": "dist/index.js",
"types": "dist/index.d.ts",
@@ -140,15 +140,15 @@
},
"dependencies": {
"@modelcontextprotocol/sdk": "^1.20.1",
"@n8n/n8n-nodes-langchain": "^1.117.0",
"@n8n/n8n-nodes-langchain": "^1.115.1",
"@supabase/supabase-js": "^2.57.4",
"dotenv": "^16.5.0",
"express": "^5.1.0",
"express-rate-limit": "^7.1.5",
"lru-cache": "^11.2.1",
"n8n": "^1.118.1",
"n8n-core": "^1.117.0",
"n8n-workflow": "^1.115.0",
"n8n": "^1.116.2",
"n8n-core": "^1.115.1",
"n8n-workflow": "^1.113.0",
"openai": "^4.77.0",
"sql.js": "^1.13.0",
"tslib": "^2.6.2",

View File

@@ -1,6 +1,6 @@
{
"name": "n8n-mcp-runtime",
"version": "2.22.8",
"version": "2.22.5",
"description": "n8n MCP Server Runtime Dependencies Only",
"private": true,
"dependencies": {

View File

@@ -1,45 +0,0 @@
#!/usr/bin/env node
/**
* Generate release notes for the initial release
* Used by GitHub Actions when no previous tag exists
*/
const { execSync } = require('child_process');
function generateInitialReleaseNotes(version) {
try {
// Get total commit count
const commitCount = execSync('git rev-list --count HEAD', { encoding: 'utf8' }).trim();
// Generate release notes
const releaseNotes = [
'### 🎉 Initial Release',
'',
`This is the initial release of n8n-mcp v${version}.`,
'',
'---',
'',
'**Release Statistics:**',
`- Commit count: ${commitCount}`,
'- First release setup'
];
return releaseNotes.join('\n');
} catch (error) {
console.error(`Error generating initial release notes: ${error.message}`);
return `Failed to generate initial release notes: ${error.message}`;
}
}
// Parse command line arguments
const version = process.argv[2];
if (!version) {
console.error('Usage: generate-initial-release-notes.js <version>');
process.exit(1);
}
const releaseNotes = generateInitialReleaseNotes(version);
console.log(releaseNotes);

View File

@@ -1,99 +0,0 @@
#!/usr/bin/env ts-node
import * as fs from 'fs';
import * as path from 'path';
import { createDatabaseAdapter } from '../src/database/database-adapter';
interface BatchResponse {
id: string;
custom_id: string;
response: {
status_code: number;
body: {
choices: Array<{
message: {
content: string;
};
}>;
};
};
error: any;
}
async function processBatchMetadata(batchFile: string) {
console.log(`📥 Processing batch file: ${batchFile}`);
// Read the JSONL file
const content = fs.readFileSync(batchFile, 'utf-8');
const lines = content.trim().split('\n');
console.log(`📊 Found ${lines.length} batch responses`);
// Initialize database
const db = await createDatabaseAdapter('./data/nodes.db');
let updated = 0;
let skipped = 0;
let errors = 0;
for (const line of lines) {
try {
const response: BatchResponse = JSON.parse(line);
// Extract template ID from custom_id (format: "template-9100")
const templateId = parseInt(response.custom_id.replace('template-', ''));
// Check for errors
if (response.error || response.response.status_code !== 200) {
console.warn(`⚠️ Template ${templateId}: API error`, response.error);
errors++;
continue;
}
// Extract metadata from response
const metadataJson = response.response.body.choices[0].message.content;
// Validate it's valid JSON
JSON.parse(metadataJson); // Will throw if invalid
// Update database
const stmt = db.prepare(`
UPDATE templates
SET metadata_json = ?
WHERE id = ?
`);
stmt.run(metadataJson, templateId);
updated++;
console.log(`✅ Template ${templateId}: Updated metadata`);
} catch (error: any) {
console.error(`❌ Error processing line:`, error.message);
errors++;
}
}
// Close database
if ('close' in db && typeof db.close === 'function') {
db.close();
}
console.log(`\n📈 Summary:`);
console.log(` - Updated: ${updated}`);
console.log(` - Skipped: ${skipped}`);
console.log(` - Errors: ${errors}`);
console.log(` - Total: ${lines.length}`);
}
// Main
const batchFile = process.argv[2] || '/Users/romualdczlonkowski/Pliki/n8n-mcp/n8n-mcp/docs/batch_68fff7242850819091cfed64f10fb6b4_output.jsonl';
processBatchMetadata(batchFile)
.then(() => {
console.log('\n✅ Batch processing complete!');
process.exit(0);
})
.catch((error) => {
console.error('\n❌ Batch processing failed:', error);
process.exit(1);
});

View File

@@ -186,115 +186,7 @@ Please choose a different name.
- Simply rename nodes with updateNode - no manual connection operations needed
- Multiple renames in one call work atomically
- Can rename a node and add/remove connections using the new name in the same batch
- Use \`validateOnly: true\` to preview effects before applying
## Removing Properties with undefined
To remove a property from a node, set its value to \`undefined\` in the updates object. This is essential when migrating from deprecated properties or cleaning up optional configuration fields.
### Why Use undefined?
- **Property removal vs. null**: Setting a property to \`undefined\` removes it completely from the node object, while \`null\` sets the property to a null value
- **Validation constraints**: Some properties are mutually exclusive (e.g., \`continueOnFail\` and \`onError\`). Simply setting one without removing the other will fail validation
- **Deprecated property migration**: When n8n deprecates properties, you must remove the old property before the new one will work
### Basic Property Removal
\`\`\`javascript
// Remove error handling configuration
n8n_update_partial_workflow({
id: "wf_123",
operations: [{
type: "updateNode",
nodeName: "HTTP Request",
updates: { onError: undefined }
}]
});
// Remove disabled flag
n8n_update_partial_workflow({
id: "wf_456",
operations: [{
type: "updateNode",
nodeId: "node_abc",
updates: { disabled: undefined }
}]
});
\`\`\`
### Nested Property Removal
Use dot notation to remove nested properties:
\`\`\`javascript
// Remove nested parameter
n8n_update_partial_workflow({
id: "wf_789",
operations: [{
type: "updateNode",
nodeName: "API Request",
updates: { "parameters.authentication": undefined }
}]
});
// Remove entire array property
n8n_update_partial_workflow({
id: "wf_012",
operations: [{
type: "updateNode",
nodeName: "HTTP Request",
updates: { "parameters.headers": undefined }
}]
});
\`\`\`
### Migrating from Deprecated Properties
Common scenario: replacing \`continueOnFail\` with \`onError\`:
\`\`\`javascript
// WRONG: Setting only the new property leaves the old one
n8n_update_partial_workflow({
id: "wf_123",
operations: [{
type: "updateNode",
nodeName: "HTTP Request",
updates: { onError: "continueErrorOutput" }
}]
});
// Error: continueOnFail and onError are mutually exclusive
// CORRECT: Remove the old property first
n8n_update_partial_workflow({
id: "wf_123",
operations: [{
type: "updateNode",
nodeName: "HTTP Request",
updates: {
continueOnFail: undefined,
onError: "continueErrorOutput"
}
}]
});
\`\`\`
### Batch Property Removal
Remove multiple properties in one operation:
\`\`\`javascript
n8n_update_partial_workflow({
id: "wf_345",
operations: [{
type: "updateNode",
nodeName: "Data Processor",
updates: {
continueOnFail: undefined,
alwaysOutputData: undefined,
"parameters.legacy_option": undefined
}
}]
});
\`\`\`
### When to Use undefined
- Removing deprecated properties during migration
- Cleaning up optional configuration flags
- Resolving mutual exclusivity validation errors
- Removing stale or unnecessary node metadata
- Simplifying node configuration`,
- Use \`validateOnly: true\` to preview effects before applying`,
parameters: {
id: { type: 'string', required: true, description: 'Workflow ID to update' },
operations: {
@@ -331,13 +223,7 @@ n8n_update_partial_workflow({
'// Vector Store setup: Connect embeddings and documents\nn8n_update_partial_workflow({id: "ai7", operations: [\n {type: "addConnection", source: "Embeddings OpenAI", target: "Pinecone Vector Store", sourceOutput: "ai_embedding"},\n {type: "addConnection", source: "Default Data Loader", target: "Pinecone Vector Store", sourceOutput: "ai_document"}\n]})',
'// Connect Vector Store Tool to AI Agent (retrieval setup)\nn8n_update_partial_workflow({id: "ai8", operations: [\n {type: "addConnection", source: "Pinecone Vector Store", target: "Vector Store Tool", sourceOutput: "ai_vectorStore"},\n {type: "addConnection", source: "Vector Store Tool", target: "AI Agent", sourceOutput: "ai_tool"}\n]})',
'// Rewire AI Agent to use different language model\nn8n_update_partial_workflow({id: "ai9", operations: [{type: "rewireConnection", source: "AI Agent", from: "OpenAI Chat Model", to: "Anthropic Chat Model", sourceOutput: "ai_languageModel"}]})',
'// Replace all AI tools for an agent\nn8n_update_partial_workflow({id: "ai10", operations: [\n {type: "removeConnection", source: "Old Tool 1", target: "AI Agent", sourceOutput: "ai_tool"},\n {type: "removeConnection", source: "Old Tool 2", target: "AI Agent", sourceOutput: "ai_tool"},\n {type: "addConnection", source: "New HTTP Tool", target: "AI Agent", sourceOutput: "ai_tool"},\n {type: "addConnection", source: "New Code Tool", target: "AI Agent", sourceOutput: "ai_tool"}\n]})',
'\n// ============ REMOVING PROPERTIES EXAMPLES ============',
'// Remove a simple property\nn8n_update_partial_workflow({id: "rm1", operations: [{type: "updateNode", nodeName: "HTTP Request", updates: {onError: undefined}}]})',
'// Migrate from deprecated continueOnFail to onError\nn8n_update_partial_workflow({id: "rm2", operations: [{type: "updateNode", nodeName: "HTTP Request", updates: {continueOnFail: undefined, onError: "continueErrorOutput"}}]})',
'// Remove nested property\nn8n_update_partial_workflow({id: "rm3", operations: [{type: "updateNode", nodeName: "API Request", updates: {"parameters.authentication": undefined}}]})',
'// Remove multiple properties\nn8n_update_partial_workflow({id: "rm4", operations: [{type: "updateNode", nodeName: "Data Processor", updates: {continueOnFail: undefined, alwaysOutputData: undefined, "parameters.legacy_option": undefined}}]})',
'// Remove entire array property\nn8n_update_partial_workflow({id: "rm5", operations: [{type: "updateNode", nodeName: "HTTP Request", updates: {"parameters.headers": undefined}}]})'
'// Replace all AI tools for an agent\nn8n_update_partial_workflow({id: "ai10", operations: [\n {type: "removeConnection", source: "Old Tool 1", target: "AI Agent", sourceOutput: "ai_tool"},\n {type: "removeConnection", source: "Old Tool 2", target: "AI Agent", sourceOutput: "ai_tool"},\n {type: "addConnection", source: "New HTTP Tool", target: "AI Agent", sourceOutput: "ai_tool"},\n {type: "addConnection", source: "New Code Tool", target: "AI Agent", sourceOutput: "ai_tool"}\n]})'
],
useCases: [
'Rewire connections when replacing nodes',
@@ -373,11 +259,7 @@ n8n_update_partial_workflow({
'Connect language model BEFORE adding AI Agent to ensure validation passes',
'Use targetIndex for fallback models (primary=0, fallback=1)',
'Batch AI component connections in a single operation for atomicity',
'Validate AI workflows after connection changes to catch configuration errors',
'To remove properties, set them to undefined (not null) in the updates object',
'When migrating from deprecated properties, remove the old property and add the new one in the same operation',
'Use undefined to resolve mutual exclusivity validation errors between properties',
'Batch multiple property removals in a single updateNode operation for efficiency'
'Validate AI workflows after connection changes to catch configuration errors'
],
pitfalls: [
'**REQUIRES N8N_API_URL and N8N_API_KEY environment variables** - will not work without n8n API access',
@@ -397,12 +279,7 @@ n8n_update_partial_workflow({
'**Auto-sanitization behavior**: Binary operators (equals, contains) automatically have singleValue removed; unary operators (isEmpty, isNotEmpty) automatically get singleValue:true added',
'**Auto-sanitization runs on ALL nodes**: When ANY update is made, ALL nodes in the workflow are sanitized (not just modified ones)',
'**Auto-sanitization cannot fix everything**: It fixes operator structures and missing metadata, but cannot fix broken connections or branch mismatches',
'**Corrupted workflows beyond repair**: Workflows in paradoxical states (API returns corrupt, API rejects updates) cannot be fixed via API - must be recreated',
'Setting a property to null does NOT remove it - use undefined instead',
'When properties are mutually exclusive (e.g., continueOnFail and onError), setting only the new property will fail - you must remove the old one with undefined',
'Removing a required property may cause validation errors - check node documentation first',
'Nested property removal with dot notation only removes the specific nested field, not the entire parent object',
'Array index notation (e.g., "parameters.headers[0]") is not supported - remove the entire array property instead'
'**Corrupted workflows beyond repair**: Workflows in paradoxical states (API returns corrupt, API rejects updates) cannot be fixed via API - must be recreated'
],
relatedTools: ['n8n_update_full_workflow', 'n8n_get_workflow', 'validate_workflow', 'tools_documentation']
}

View File

@@ -75,15 +75,10 @@ async function fetchTemplatesRobust() {
// Fetch detail
const detail = await fetcher.fetchTemplateDetail(template.id);
if (detail !== null) {
// Save immediately
repository.saveTemplate(template, detail);
saved++;
} else {
errors++;
console.error(`\n❌ Failed to fetch template ${template.id} (${template.name}) after retries`);
}
// Save immediately
repository.saveTemplate(template, detail);
saved++;
// Rate limiting
await new Promise(resolve => setTimeout(resolve, 200));

View File

@@ -40,37 +40,7 @@ export interface TemplateDetail {
export class TemplateFetcher {
private readonly baseUrl = 'https://api.n8n.io/api/templates';
private readonly pageSize = 250; // Maximum allowed by API
private readonly maxRetries = 3;
private readonly retryDelay = 1000; // 1 second base delay
/**
* Retry helper for API calls
*/
private async retryWithBackoff<T>(
fn: () => Promise<T>,
context: string,
maxRetries: number = this.maxRetries
): Promise<T | null> {
let lastError: any;
for (let attempt = 1; attempt <= maxRetries; attempt++) {
try {
return await fn();
} catch (error: any) {
lastError = error;
if (attempt < maxRetries) {
const delay = this.retryDelay * attempt; // Exponential backoff
logger.warn(`${context} - Attempt ${attempt}/${maxRetries} failed, retrying in ${delay}ms...`);
await this.sleep(delay);
}
}
}
logger.error(`${context} - All ${maxRetries} attempts failed, skipping`, lastError);
return null;
}
/**
* Fetch all templates and filter to last 12 months
* This fetches ALL pages first, then applies date filter locally
@@ -103,105 +73,93 @@ export class TemplateFetcher {
let page = 1;
let hasMore = true;
let totalWorkflows = 0;
logger.info('Starting complete template fetch from n8n.io API');
while (hasMore) {
const result = await this.retryWithBackoff(
async () => {
const response = await axios.get(`${this.baseUrl}/search`, {
params: {
page,
rows: this.pageSize
// Note: sort_by parameter doesn't work, templates come in popularity order
}
});
return response.data;
},
`Fetching templates page ${page}`
);
if (result === null) {
// All retries failed for this page, skip it and continue
logger.warn(`Skipping page ${page} after ${this.maxRetries} failed attempts`);
try {
const response = await axios.get(`${this.baseUrl}/search`, {
params: {
page,
rows: this.pageSize
// Note: sort_by parameter doesn't work, templates come in popularity order
}
});
const { workflows } = response.data;
totalWorkflows = response.data.totalWorkflows || totalWorkflows;
allTemplates.push(...workflows);
// Calculate total pages for better progress reporting
const totalPages = Math.ceil(totalWorkflows / this.pageSize);
if (progressCallback) {
// Enhanced progress with page information
progressCallback(allTemplates.length, totalWorkflows);
}
logger.debug(`Fetched page ${page}/${totalPages}: ${workflows.length} templates (total so far: ${allTemplates.length}/${totalWorkflows})`);
// Check if there are more pages
if (workflows.length < this.pageSize) {
hasMore = false;
}
page++;
continue;
}
const { workflows } = result;
totalWorkflows = result.totalWorkflows || totalWorkflows;
allTemplates.push(...workflows);
// Calculate total pages for better progress reporting
const totalPages = Math.ceil(totalWorkflows / this.pageSize);
if (progressCallback) {
// Enhanced progress with page information
progressCallback(allTemplates.length, totalWorkflows);
}
logger.debug(`Fetched page ${page}/${totalPages}: ${workflows.length} templates (total so far: ${allTemplates.length}/${totalWorkflows})`);
// Check if there are more pages
if (workflows.length < this.pageSize) {
hasMore = false;
}
page++;
// Rate limiting - be nice to the API (slightly faster with 250 rows/page)
if (hasMore) {
await this.sleep(300); // 300ms between requests (was 500ms with 100 rows)
// Rate limiting - be nice to the API (slightly faster with 250 rows/page)
if (hasMore) {
await this.sleep(300); // 300ms between requests (was 500ms with 100 rows)
}
} catch (error) {
logger.error(`Error fetching templates page ${page}:`, error);
throw error;
}
}
logger.info(`Fetched all ${allTemplates.length} templates from n8n.io`);
return allTemplates;
}
async fetchTemplateDetail(workflowId: number): Promise<TemplateDetail | null> {
const result = await this.retryWithBackoff(
async () => {
const response = await axios.get(`${this.baseUrl}/workflows/${workflowId}`);
return response.data.workflow;
},
`Fetching template detail for workflow ${workflowId}`
);
return result;
async fetchTemplateDetail(workflowId: number): Promise<TemplateDetail> {
try {
const response = await axios.get(`${this.baseUrl}/workflows/${workflowId}`);
return response.data.workflow;
} catch (error) {
logger.error(`Error fetching template detail for ${workflowId}:`, error);
throw error;
}
}
async fetchAllTemplateDetails(
workflows: TemplateWorkflow[],
workflows: TemplateWorkflow[],
progressCallback?: (current: number, total: number) => void
): Promise<Map<number, TemplateDetail>> {
const details = new Map<number, TemplateDetail>();
let skipped = 0;
logger.info(`Fetching details for ${workflows.length} templates`);
for (let i = 0; i < workflows.length; i++) {
const workflow = workflows[i];
const detail = await this.fetchTemplateDetail(workflow.id);
if (detail !== null) {
try {
const detail = await this.fetchTemplateDetail(workflow.id);
details.set(workflow.id, detail);
} else {
skipped++;
logger.warn(`Skipped workflow ${workflow.id} after ${this.maxRetries} failed attempts`);
if (progressCallback) {
progressCallback(i + 1, workflows.length);
}
// Rate limiting (conservative to avoid API throttling)
await this.sleep(150); // 150ms between requests
} catch (error) {
logger.error(`Failed to fetch details for workflow ${workflow.id}:`, error);
// Continue with other templates
}
if (progressCallback) {
progressCallback(i + 1, workflows.length);
}
// Rate limiting (conservative to avoid API throttling)
await this.sleep(150); // 150ms between requests
}
logger.info(`Successfully fetched ${details.size} template details (${skipped} skipped)`);
logger.info(`Successfully fetched ${details.size} template details`);
return details;
}

View File

@@ -496,17 +496,10 @@ export class TemplateRepository {
// Count node usage
const nodeCount: Record<string, number> = {};
topNodes.forEach(t => {
if (!t.nodes_used) return;
try {
const nodes = JSON.parse(t.nodes_used);
if (Array.isArray(nodes)) {
nodes.forEach((n: string) => {
nodeCount[n] = (nodeCount[n] || 0) + 1;
});
}
} catch (error) {
logger.warn(`Failed to parse nodes_used for template stats:`, error);
}
const nodes = JSON.parse(t.nodes_used);
nodes.forEach((n: string) => {
nodeCount[n] = (nodeCount[n] || 0) + 1;
});
});
// Get top 10 most used nodes