n8n AI Agent Tools List

n8n AI Agent Tools List

The power of n8n's AI Agent node comes from its ability to use tools — external functions the agent can call to fetch data, perform actions, or compute results. This guide catalogs every built-in tool, shows how to configure each one, and provides patterns for building custom tools.

How Tools Work in n8n AI Agent

When you add a tool to the AI Agent node:

1. Agent decides to use a tool based on user query 2. Agent prepares parameters (validated against tool schema) 3. n8n executes the tool node 4. Result returned to agent as context 5. Agent responds to user with tool output

Tool Schema: Each tool exposes a JSON Schema defining its name, description, and parameters. The LLM uses this to decide when and how to call the tool.

Built-in Tool Categories

1. Web Search Tools

SerpAPI (Google Search)

Setup: 1. Get API key from serpapi.com 2. Add HTTP Request node 3. Method: GET 4. URL: https://serpapi.com/search 5. Query Parameters: q={{$json.query}}, api_key={{$credentials.serpApiKey}}, num=10 6. AI Agent Tool Config: - Name: web_search - Description: Search Google for current information, news, facts - Parameters: { query: { type: string, description: 'Search query' } }

Brave Search API

DuckDuckGo Instant Answer (Free)

2. Database Tools

PostgreSQL

Common Patterns: - Select: SELECT * FROM users WHERE email = $1 - Insert: INSERT INTO leads (email, name) VALUES ($1, $2) RETURNING id - Update: UPDATE users SET last_login = NOW() WHERE id = $1

Security: Use parameterized queries ($1, $2) — never interpolate user input directly.

MySQL / MariaDB

MongoDB

Redis

Supabase (PostgreSQL + Auth + Realtime)

3. API & HTTP Tools

Generic HTTP Request

Common Patterns: - REST API: GET/POST/PUT/DELETE with JSON body - GraphQL: POST with query/variables in body - Webhooks: POST with signature verification - File Upload: Multipart form data

REST API Best Practices

// Tool Schema for Generic API Call
{
  "name": "api_call",
  "description": "Make HTTP request to any REST API",
  "parameters": {
    "type": "object",
    "properties": {
      "method": { "type": "string", "enum": ["GET", "POST", "PUT", "DELETE"] },
      "url": { "type": "string", "format": "uri" },
      "headers": { "type": "object" },
      "body": { "type": "object" },
      "timeout": { "type": "number", "default": 30000 }
    },
    "required": ["method", "url"]
  }
}

GraphQL

4. Calculation & Computation Tools

Calculator

// Tool: calculator
const { expression } = $json;
try {
  // Safe eval — only allow math
  const result = Function('"use strict"; return (' + expression + ')')();
  return { result };
} catch (e) {
  return { error: 'Invalid expression' };
}

Date/Time Calculator

// Tool: date_calculator
const { operation, date, value, unit, timezone } = $json;
const dateObj = new Date(date);
switch (operation) {
  case 'add': dateObj.setDate(dateObj.getDate() + value); break;
  case 'subtract': dateObj.setDate(dateObj.getDate() - value); break;
  case 'format': return { result: format(dateObj, value, { timeZone: timezone }) };
  case 'diff': return { result: differenceInDays(new Date(), dateObj) };
}
return { result: dateObj.toISOString() };

Unit Converter

5. Code Execution Tools

JavaScript / TypeScript Execution

Security: Code runs in isolated VM. No filesystem/network access unless explicitly allowed.

// Tool: code_executor
const { code, language, timeout } = $json;
// Language: javascript (only supported)
// Timeout in ms (max 300000)
const start = Date.now();
try {
  const fn = new Function('return ' + code);
  const result = fn();
  return { result, executionTime: Date.now() - start };
} catch (e) {
  return { error: e.message };
}

Python Execution (via HTTP to external service)

6. Communication Tools

Email (SendGrid, SMTP, Gmail, Outlook)

SMS (Twilio, Vonage, MessageBird, Plivo)

Slack

Microsoft Teams

Discord

WhatsApp (Twilio, Vonage, MessageBird, 360Dialog)

Telegram

7. File & Document Tools

PDF Processing

Document Generation

File Storage

CSV / Excel Processing

8. AI & ML Tools

OpenAI / Anthropic / Google AI

Hugging Face

Local LLMs (Ollama, LM Studio, vLLM)

9. Workflow & Automation Tools

Workflow Control

Data Transformation

10. Custom Tool Development

Creating Custom Tools

1. As a Workflow (Reusable)

1. Create workflow: tools/my_custom_tool 2. Input: Webhook with defined schema 3. Logic: Any n8n nodes 4. Output: Respond to Webhook with result 5. Register as Tool: In AI Agent node > Tools > Add Custom Workflow

2. As a Custom Node (Advanced)

export class MyCustomTool implements INodeType { description: INodeTypeDescription = { displayName: 'My Custom Tool', name: 'myCustomTool', group: ['transform'], version: 1, description: 'Custom tool for AI Agent', defaults: { name: 'My Custom Tool' }, inputs: ['main'], outputs: ['main'], credentials: [], properties: [ { displayName: 'Input', name: 'input', type: 'string', default: '', description: 'Input for the tool', }, ], };

async execute(this: IExecuteFunctions): Promise { const input = this.getNodeParameter('input', 0); // Your logic here return this.prepareOutputData([{ json: { result: Processed: ${input} } }]); } } ```

Tool Selection Guide

Tool Configuration Best Practices

1. Name clearly: web_search, database_query, send_email 2. Describe precisely: "Search Google for current information" not "Search" 3. Parameter descriptions: Help LLM understand each field 4. Validate inputs: Use schema validation, required fields 5. Handle errors: Return structured error objects 6. Limit scope: One tool = one clear purpose 7. Test schemas: Use n8n's AI Agent test panel

Tool Limits & Quotas

Related Guides

• n8n AI Agent Node Tutorial - Agent basics
• n8n AI Agent Memory Setup - Memory configuration
• n8n RAG Agent Tutorial - RAG with tools
• n8n AI Agent Error Handling - Error patterns
• n8n Google Sheets Automation - Sheets as tool