Generate SQL Queries from Database Schema with n8n AI Agent

What this workflow does

This workflow helps you ask questions in simple language and get correct SQL queries run on your MySQL database automatically.
It fixes the hard task of writing SQL by yourself and quickly gives you data answers.

You send a chat message like “Show all customers from Germany”.
The workflow uses AI with your database structure to make a SQL query.
It runs this query and sends the results back.

Who should use this workflow

This is great for people who do not know SQL very well.
Anyone who needs fast answers from a MySQL database without typing complex queries.

It helps data analysts, business teams, or anyone new to databases.
It saves time and avoids mistakes from manual query writing.

Tools and services used

• MySQL database connection: To get schema and run queries.
• OpenAI API with GPT-4: Converts your words to SQL queries.
• n8n automation platform: To build workflow nodes.
• LangChain Chat nodes: For AI chat trigger and AI agent processing.

Use self-host n8n if you want full control.

Inputs, processing steps, and output explained

Inputs

• User sends a natural language question via chat.
• Workflow has database schema loaded from a local JSON file.

Processing steps

• Load database tables and schema once to avoid delays.
• Receive chat input in natural language.
• Combine the user question with the saved schema.
• Use GPT-4 AI to create a matching SQL query.
• Extract and verify that the SQL query is valid.
• Run the SQL on MySQL database if valid.
• Format results as a readable Markdown table.
• Send back both chat answer and data results.

Output

The user gets an easy answer to their question plus actual data in clear table form.
This speeds up reports and prevents query mistakes.

Beginner step-by-step: How to use this workflow in production inside n8n

Importing the workflow

1. Download the full workflow file using the Download button on this page.
2. Open the n8n editor where the automation runs.
3. Click “Import from File” and select the downloaded workflow.

Configuring the workflow

1. Add your MySQL database credentials inside all MySQL node settings.
2. Enter your OpenAI API Key in the LangChain AI Agent node credentials.
3. Check table names and schema file paths if your database is different.

Testing and activating

1. Run a test chat input by the LangChain Chat Trigger node with a sample question like “Show German customers”.
2. Verify the workflow creates a SQL query and returns data.
3. If successful, activate the workflow.
4. Make sure the chat trigger webhook is public or connected to your chat system.

Now, you can ask your database questions in plain words.

Customization ideas

• Switch database types by changing credentials and queries in all MySQL nodes to PostgreSQL or other SQL engines.
• Modify the AI Agent’s prompt in the LangChain AI Agent node to make it answer differently or use other languages.
• Automate refreshing the JSON schema file regularly to keep data fresh.
• Use encrypted storage for the schema file to keep it secure.
• Add support for multiple databases by changing the schema extraction and AI prompt.

Dealing with common problems

• MySQL connection errors: Check username, password, host, port in MySQL nodes.
• No SQL query generated: Make input clearer or update the AI prompt in LangChain AI Agent node.
• Regex does not catch SQL query: Edit regex in Set node ‘Extract SQL query’ to match AI response.
• Slow response: Confirm schema is loading from local file to avoid repeated slow database schema calls.

Pre-production checklist

• Test database connection returns tables.
• Verify schema extraction gathers correct columns.
• Confirm chat trigger receives questions and AI creates SQL.
• Check saved local JSON file for schema exists.
• Validate AI Agent makes valid SQL queries.
• Test workflow branching with and without queries.
• Backup workflow and database before live use.

Deployment guide

Activate the workflow in n8n once setup and tested.
Make sure the LangChain Chat Trigger node webhook is accessible from your chat app.

Watch workflow executions panel to catch errors.
Set alerts for database or OpenAI API issues if possible.

Summary: What you achieve

✓ Save hours by not writing SQL manually.
✓ Get quick, clear data answers from your MySQL.
✓ Avoid SQL mistakes through AI validation.
✓ Help non-technical users ask questions naturally.
✓ Keep database schema handy for fast AI query generation.

→ Use the workflow in real chats for fast, accurate data insights.
→ Customize AI behavior and database connections as needed.

Code and prompt sample for AI Agent (important)

The LangChain AI Agent uses your schema and user input combined in this JSON format:

{
 "sessionId": "{{sessionId}}",
 "action": "{{action}}",
 "chatinput": "{{chatinput}}",
 "schema": "{{schemaJsonString}}"
}

This lets GPT-4 know exactly what tables and columns exist before making SQL.
A regex in the Set node ‘Extract SQL query’ looks for SELECT ... ; statements in the AI’s reply.
Use this to be sure only valid queries run on the database.