🔄 n8n MCP Server Configuration

Complete guide to integrate n8n with AI4CALL: connect your assistant to the powerful open-source workflow automation platform via webhook

Go to "MARKET" → "Marketplace Plugin" and click Purchase for "n8n" → FREE

Go to "MARKET" → "My Plugins" and click Configure for "n8n"

🚀 Quick Navigation:

⚠️ Third-party services

The integrations described refer to third-party services.
Rate limits, authentication methods, technical configurations, costs, and policies are defined exclusively by the respective providers and may change over time.

Costs for external services are borne directly by the end user to the service providers themselves and are in no way managed or mediated by our portal (e.g., Meta or other providers for WhatsApp, Zapier, Calendly, etc.).

💡 n8n vs Zapier: which to choose?

n8n - open-source self-hosted workflow automation:

Advantages:

✅ no monthly task/execution limits
✅ complete data control (privacy, GDPR)
✅ complex workflow logic with JavaScript
✅ direct connection to SQL databases
✅ predictable costs (only server hosting)
✅ unlimited customization

Disadvantages:

⚠️ requires server setup (Docker, cloud VM, or n8n.cloud)
⚠️ fewer pre-integrated apps vs Zapier (but HTTP Request covers 90%)
⚠️ steeper learning curve

When to use n8n: high call volume, complex logic, data control, limited long-term budget
When to use Zapier: quick setup, many pre-integrated apps, no server management

What is n8n and how the integration works

n8n (node to node) is an open-source workflow automation platform that allows connecting apps and services via visual drag-and-drop interface.

🏗️ n8n architecture:

Workflow:
→ each node = an operation (read file, call API, send email, query database)

Trigger:
→ timer (cron), HTTP webhook, file/database monitoring, external events

Action Nodes:
→ HTTP Request, Database Query, Email, Slack, Google Sheets, IF/Switch logic

Execution:
→ visible in real time with detailed debugging for each node

🔌 How it integrates with AI4CALL

Integration architecture:


[Customer calls AI4CALL]
   ↓
[AI Assistant collects data]
   ↓
[AI formats data into JSON]
   ↓
[POST to n8n webhook]
   ↓
[n8n Workflow activates]
   ↓
[n8n Nodes process data]
   ↓
[Actions executed (CRM, Email, DB, etc.)]
   ↓
[n8n returns response (optional)]

Key difference vs Zapier: n8n gives you total control over workflow logic - you can write custom JavaScript, make direct SQL queries, implement complex retry logic, all visually debuggable.

💻 n8n deployment options

1. n8n.cloud (official hosted):

✅ immediate setup, no server management
✅ free plan available (5,000 workflow executions/month)
✅ webhook URL already configured: https://[instance].app.n8n.cloud/webhook/[path]
⚠️ costs grow with executions (similar to Zapier)

2. Self-hosted Docker (recommended):

✅ complete control, no execution limits
✅ fixed costs (only server: $5-20/month VPS)
✅ total data privacy
⚠️ requires Linux server with Docker installed

3. Cloud providers (AWS, Azure, GCP):

✅ enterprise scalability
✅ high reliability
⚠️ more complex setup, variable costs

To start: use n8n.cloud free plan. For production: consider self-hosted on economical VPS.

Create n8n workflow with Webhook node

To receive data from AI4CALL, you need to create an n8n workflow that starts with a correctly configured Webhook node.

🎯 procedure on n8n:

Access your n8n instance (n8n.cloud or self-hosted)
Click "New workflow" (or "+" at top)
In the node palette on left, search for "Webhook"
Drag the "Webhook" node onto the canvas
Click on the Webhook node to configure it:
- HTTP Method: select POST
- Path: enter custom path (e.g., ai4call-integration)
- Authentication: choose type (None, Basic Auth, Header Auth)
- Response Mode: "On last node" for synchronous response, "Webhook node" for immediate
n8n will show the generated webhook URL:
https://your-n8n-instance.com/webhook/ai4call-integration
Copy this URL (you will use it in AI4CALL)
Add subsequent nodes to the workflow to process received data
SAVE the workflow (Ctrl+S or Save button)
ACTIVATE the workflow (toggle ON top right)

💡 Basic workflow example: save data to Google Sheets

n8n workflow in 3 nodes:

Webhook: receives data from AI4CALL (POST /webhook/ai4call)
→ payload: {name, email, phone, notes}
Google Sheets: add row to sheet
- Map fields: {{$json.name}} → Name column
- Map fields: {{$json.email}} → Email column
- etc.
Respond to Webhook: confirm success
→ response: {"status": "success", "row_id": "..."}

Save, activate, and the workflow is ready to receive data from AI4CALL! 🎉

⚠️ Production URL vs Test URL

FUNDAMENTAL: always use the Production URL, never the Test URL!

n8n generates TWO URLs for each webhook:

Test URL: https://.../webhook-test/...

❌ temporary - deactivates after timeout
❌ only for manual testing in n8n
❌ NOT for persistent integrations

Production URL: https://.../webhook/...

✅ permanent as long as workflow is active
✅ use this for AI4CALL
✅ requires workflow saved and activated (toggle ON)

Always copy the Production URL into the "Webhook URL" field in AI4CALL!

Configure the webhook in AI4CALL

Insert the n8n webhook URL in the AI4CALL configuration to allow the assistant to send data to your workflows.

🎯 procedure on AI4CALL:

Select "n8n" from your Plugin list
In the "Webhook URL" field, paste the complete Production URL copied from n8n
Verify the URL is complete:
- Must start with https://
- Must contain /webhook/ (NOT /webhook-test/)
- Must include the path configured in n8n
Proceed with authentication configuration (see next step)
Click "SAVE CONFIGURATION"

📋 Correct URL format

n8n.cloud:


https://[your-instance].app.n8n.cloud/webhook/[your-path]

Example: https://johndoe.app.n8n.cloud/webhook/ai4call-integration

n8n self-hosted:


https://[your-domain-or-ip]/webhook/[your-path]

Example: https://n8n.mycompany.com/webhook/ai4call-integration
or: http://123.45.67.89:5678/webhook/ai4call-integration (if HTTP)

⚠️ Self-hosted: if using HTTP (no HTTPS), ensure AI4CALL accepts insecure connections, or better configure HTTPS with Let's Encrypt/Cloudflare.

🔍 Verify webhook accessibility

Before saving in AI4CALL, test that the webhook is reachable:

Test with curl:


curl -X POST https://your-n8n-instance.com/webhook/ai4call-integration \
  -H "Content-Type: application/json" \
  -d '{"test": "data"}'

Expected response: HTTP 200 OK with JSON response

If you get an error:

❌ 404 Not Found → verify URL and webhook path
❌ 401/403 Unauthorized → authentication problem
❌ Connection refused → workflow not active or server unreachable
❌ Timeout → firewall blocks traffic or n8n server down

"Configure n8n" panel: webhook URL entry with example and authentication type instructions

Configure webhook authentication

Protect the n8n webhook by configuring the appropriate authentication type both in n8n and in AI4CALL.

🔒 available authentication types:

1. No Authentication

When to use: public webhook, testing, demo
Security: ❌ anyone with the URL can send data
n8n configuration: Authentication = None
AI4CALL configuration: Authentication Type = "No Authentication"

2. Bearer Token

When to use: standard authentication, common REST APIs
Security: ✅ requires secret token in Authorization header
n8n configuration: Authentication = Header Auth
→ Header Name: Authorization
→ Header Value: Bearer YOUR_SECRET_TOKEN
AI4CALL configuration: Authentication Type = "Bearer Token"
→ Token: YOUR_SECRET_TOKEN (without "Bearer " prefix)

3. Header Auth (custom)

When to use: custom authentication, legacy API
Security: ✅ custom HTTP header with secret value
n8n configuration: Authentication = Header Auth
→ Header Name: X-API-Key (or custom name)
→ Header Value: your-secret-api-key
AI4CALL configuration: Authentication Type = "Header Auth"
→ Header Name: X-API-Key
→ Header Value: your-secret-api-key

✅ Recommended configuration: Bearer Token

For most cases, use Bearer Token:

Steps on n8n:

In the Webhook node, Authentication = Header Auth
Header Name: Authorization
Header Value: Bearer sk_live_abc123xyz789 (generate long random token)
Save workflow

Steps on AI4CALL:

Authentication Type: select "Bearer Token"
Token: enter sk_live_abc123xyz789 (WITHOUT "Bearer " prefix)
Save configuration

AI4CALL will automatically send: Authorization: Bearer sk_live_abc123xyz789

🔐 Generate secure tokens

Do NOT use trivial tokens like "secret123" or "mytoken"!

Generate long random tokens (32+ characters) with:

Method 1 - Online: https://randomkeygen.com → 504-bit WPA Key
Method 2 - Linux/Mac: openssl rand -hex 32
Method 3 - Node.js: require('crypto').randomBytes(32).toString('hex')

Example secure token: 7f3a9b2c8d1e4f5g6h7i8j9k0l1m2n3o4p5q6r7s8t9u0v1w2x3y4z5

Store the token securely (password manager) and do NOT commit it to Git!

Instruct the assistant to use n8n

Configure the assistant prompt to correctly use n8n webhooks in automatic workflows.

📝 Recommended workflows to insert in the prompt

1. Automatic CRM update:
"When customer provides new contact data (email, phone, address), immediately update CRM using n8n webhook. Send: {customer_id, field_name, new_value, timestamp}. n8n workflow will update HubSpot/Salesforce with validation."

2. Automatic support ticket:
"If customer reports technical problem or assistance request, automatically create Jira/Zendesk ticket via n8n. Send: {customer_name, email, issue_type, description, priority}. Confirm to customer the ticket number received from n8n response."

3. Real-time team notifications:
"For qualified lead or critical urgency, send immediate notification to Slack/Teams via n8n. Use dedicated n8n webhook for notifications. Format: {alert_type, customer_info, urgency_level, action_required}."

4. Database synchronization:
"After each booked appointment, synchronize with company database via n8n. n8n workflow will execute SQL INSERT/UPDATE queries. Handle error if database unavailable and inform customer."

🧪 System test

Verify end-to-end integration:

Call the AI4CALL assistant
Trigger a workflow that should invoke n8n (e.g., "update my email")
The assistant should send data to the n8n webhook
On n8n, go to Executions and verify the workflow executed
Check received payload and output of each node
Verify on final system (CRM, DB, etc.) that data arrived

Debugging:

n8n Executions Log shows received payload, node outputs, detailed errors
If workflow doesn't activate: verify URL, authentication, workflow ON
If workflow fails: check error message in red node
Use "Edit Fields" node in n8n for data transformation debugging

🛠️ n8n Webhook System

Technical specifications on how the AI assistant communicates with n8n via HTTP POST webhook and JSON data format.

Communication architecture

The AI assistant sends data to n8n via HTTP POST request with JSON payload and authentication header.

📡 HTTP request format:


POST https://your-n8n.com/webhook/ai4call-integration
Content-Type: application/json
Authorization: Bearer sk_live_abc123xyz789

{
  "customer_name": "Mario Rossi",
  "email": "mario.rossi@example.com",
  "phone": "+393331234567",
  "action": "update_contact",
  "timestamp": "2025-01-03T14:30:00Z"
}

✅ n8n response (synchronous):


HTTP/1.1 200 OK
Content-Type: application/json

{
  "status": "success",
  "workflow_id": "wf_12345",
  "execution_id": "exec_67890",
  "result": {
    "crm_updated": true,
    "record_id": "contact_999"
  }
}

Note: the response depends on how you configured the n8n workflow. With "Respond to Webhook" node you can completely customize the JSON response that the assistant will receive.

Common n8n workflow examples

🗄️ CRM Update (HubSpot)

n8n workflow:
1. Webhook (receives data from AI4CALL)
2. HubSpot node: Update Contact

Email: {{$json.email}}
Phone: {{$json.phone}}
Last Updated: {{$json.timestamp}}

3. Respond to Webhook: {"status": "updated"}

🎫 Create Ticket (Jira)

n8n workflow:
1. Webhook (receives data from AI4CALL)
2. IF node: check priority
3a. Priority HIGH: Jira Create Issue + Slack Alert
3b. Priority NORMAL: Jira Create Issue
4. Respond to Webhook: {"ticket_id": "PROJ-123"}

💾 Database Sync (PostgreSQL)

n8n workflow:
1. Webhook (receives data from AI4CALL)
2. Postgres node: INSERT INTO calls


INSERT INTO calls (customer, phone, notes)
VALUES ('{{$json.name}}', '{{$json.phone}}', '{{$json.notes}}')

3. Respond to Webhook: {"inserted": true}

📧 Email Automation (SMTP)

n8n workflow:
1. Webhook (receives data from AI4CALL)
2. Function node: generate HTML email template
3. Send Email node (SMTP/Gmail)

To: {{$json.customer_email}}
Subject: Confirmation {{$json.request_type}}
Body: HTML template with merge fields

4. Respond: {"email_sent": true}

❓ Frequently Asked Questions

Answers to the most common questions about n8n MCP Server integration

? What is n8n and how does it differ from Zapier?

n8n is an open-source self-hosted workflow automation platform, similar to Zapier but with complete control over infrastructure. Unlike Zapier (cloud SaaS), n8n can be installed on your own server, offering greater privacy, unlimited customization, no monthly task limits, and predictable costs. Integration with AI4CALL works via HTTP webhook.

? How do I get the n8n webhook URL?

In n8n, create a new workflow, add a "Webhook" node, configure HTTP method (POST), set the webhook path (e.g., /ai4call-webhook). n8n will automatically generate the complete URL in the format https://your-n8n-instance.com/webhook/ai4call-webhook. If using n8n.cloud, the URL will be https://[your-instance].app.n8n.cloud/webhook/[path].

? What authentication types does the n8n integration support?

Three types: (1) No Authentication - public webhook without protection, (2) Bearer Token - Authorization: Bearer YOUR_TOKEN header for standard authentication, (3) Header Auth - custom HTTP header with specific name and value for custom authentication. Choose the type corresponding to the n8n webhook configuration.

? Do I need to install n8n or can I use n8n.cloud?

Both options work with AI4CALL. n8n.cloud is the official hosted version (free plan available with limits), ideal to start quickly. Self-hosted n8n requires your own server (Docker, npm, or cloud provider) but offers total control, no task limits, and fixed costs. The webhook integration works identically in both cases.

? How do I test if the n8n webhook is configured correctly?

Method 1: use the integrated "Test Webhook" in n8n - activate test mode in the Webhook node, manually send a POST request with example JSON payload using Postman/curl. Method 2: configure the URL in AI4CALL and make a test call to the assistant asking it to send test data. Verify in n8n Execution Log that the workflow executed and check the received payload.

? Does the n8n webhook always need to be in Production mode?

Yes. n8n webhooks in "Test" mode are temporary and deactivate after timeout. For permanent AI4CALL integration, the workflow must be: (1) saved and activated (toggle ON top right), (2) the webhook set to "Production URL" not "Test URL". Otherwise AI4CALL won't be able to reach the webhook after a few minutes.

? How do I handle errors if the n8n workflow fails?

n8n has integrated error handling: (1) add "Error Trigger" node in workflow to catch errors and handle them (e.g., retry, notification, logging), (2) configure webhook response to return HTTP 4xx/5xx error that AI4CALL can interpret, (3) use n8n Execution Log for detailed debugging, (4) set automatic email/Slack notifications in n8n for failed workflows.

? What is the difference between n8n webhook and Zapier webhook?

n8n webhook: self-hosted, complete control over workflow logic, unlimited nodes, real-time visible execution, detailed debugging, advanced data transformation with JavaScript, direct database connections. Zapier webhook: cloud managed, simpler to configure, 6000+ pre-integrated apps, no server management, but limited to monthly tasks and variable costs. n8n offers more flexibility, Zapier more speed.

? Can I pass files or attachments via n8n webhook?

Not directly in the JSON payload. For files: (1) upload file to cloud storage (Google Drive, S3, etc.) using another MCP tool, (2) obtain public/temporary URL of the file, (3) pass the URL via n8n webhook, (4) in n8n workflow use HTTP Request node to download file from URL, (5) process file with specific nodes (PDF, Excel, Image, etc.).

? How do I monitor n8n workflow executions triggered by AI4CALL?

n8n Execution Log shows all executions in real time: see received payload, output of each node, errors, duration. Access from Executions menu in n8n. You can filter by workflow, status (success/error/running), time period. For advanced monitoring: export execution data, integrate with Prometheus/Grafana, configure custom alerting via n8n webhook itself.

? Does n8n support synchronous webhook responses or only asynchronous?

Both. The n8n Webhook node can respond synchronously: the workflow processes data and returns an immediate HTTP response with custom JSON payload. The AI4CALL assistant can receive this response and use it in conversation (e.g., "The system confirmed order #12345"). For long processes, use asynchronous mode: webhook immediately responds 200 OK, workflow continues in background.

? Can I protect the n8n webhook with custom authentication beyond Bearer Token?

Yes, use "Header Auth" in AI4CALL. Configure a custom HTTP header (e.g., X-API-Key: your-secret-key). In the n8n workflow, add an "If" node after Webhook to validate the header: read $request.headers["x-api-key"], compare with expected value, if no match return 401 Unauthorized. You can also implement more complex validation with JavaScript code node.

? How do I integrate n8n with SQL database directly from the workflow?

n8n has integrated database nodes: PostgreSQL, MySQL, MongoDB, SQLite, MSSQL. In workflow after Webhook: (1) add corresponding database node, (2) configure connection (host, port, user, password, database), (3) write SQL query or use visual query builder, (4) use data from webhook in WHERE/VALUES: {{$json.customer_id}}, (5) return query results in webhook response if needed.

? Can I use n8n to orchestrate multiple MCP tools in sequence?

Absolutely. Example workflow: (1) Webhook receives data from AI4CALL, (2) HTTP Request node searches customer on CRM API, (3) Switch node validates if customer exists, (4a) if exists: MySQL update database, (4b) if not exists: HTTP POST create customer, (5) Slack node sends notification, (6) Email confirmation node. n8n excels in multi-step orchestration with conditional branching.

? How do I handle rate limiting if AI4CALL sends many requests to n8n?

Self-hosted n8n has no intrinsic rate limit (depends on server resources). To handle high volume: (1) scale n8n horizontally with queue mode + Redis, (2) use custom "Rate Limiter" node with JavaScript code, (3) implement caching with Redis/database to reduce downstream load, (4) monitor n8n CPU/memory and scale container/VM if needed.

? Can I use saved credentials in n8n to access external APIs?

Yes, it's a key feature. n8n Credentials centrally manages API keys, OAuth tokens, database passwords securely (encrypted). In workflow: (1) configure credentials for service (e.g., Google, Slack, custom API), (2) in HTTP Request nodes select saved credential instead of hardcoding token, (3) n8n automatically injects correct header/auth. Improves security and maintenance.

? How do I debug when webhook receives data but workflow doesn't activate?

Verify checklist: (1) is workflow activated (toggle ON)? (2) is webhook Production not Test? (3) correct HTTP method (POST)? (4) webhook path exactly matches? (5) check Execution Log for errors, (6) verify firewall/network if self-hosted, (7) test webhook manually with curl/Postman, (8) check n8n system logs (Docker logs or PM2 logs if self-hosted).

? Does n8n support custom timeouts for long workflows?

Yes. However, in phone flows it is not recommended to exceed a few seconds of synchronous execution.

Use:
- Webhook with immediate response (ack) to the phone system
- Asynchronous workflows (Execute Workflow, code, events) for long processing
- Wait node only for short, controlled pauses

The call must never wait for the workflow.

? Can I migrate workflows from Zapier to n8n to use them with AI4CALL?

There is no automatic direct import, but you can recreate logic manually: (1) map Zapier triggers → n8n Webhook, (2) map Zapier actions → corresponding n8n nodes (e.g., Google Sheets Zapier action → Google Sheets n8n node), (3) transform Zapier filters → n8n If/Switch nodes, (4) multi-step Zaps → n8n workflow with sequential nodes. n8n is more flexible but requires initial manual configuration.