Generate SQL Queries from Schema Only — n8n AI Agent Template

Input: natural language question + local database schema file → Processing: AI Agent generates SQL based solely on schema (no data access) → Output: SQL query and execution results (or direct answer)

Who

Developers or analysts who need to query databases via natural language without exposing raw data

You get

Ask via chat window; AI generates SQL using only schema structure, executed externally with formatted results returned; non-SQL questions get direct text answers

You need

n8n · mysql

Steps

1. Prepare environment and one-time schema download

   1) Prepare a MySQL database and import data (template example uses Chinook dataset from GitHub). 2) Configure mySql and openAiApi credentials in n8n. 3) Run the top half of the workflow (Manual Trigger → List all tables → Extract database schema → Add table name → Convert to binary → Save file locally) to save schema as ./chinook_mysql.json. This step runs only once.

2. Chat trigger and load local schema

   User sends a question in the Chat Trigger window. The workflow reads the locally saved ./chinook_mysql.json, extracts it as a JSON object, and merges it with sessionId, action, and chatInput from the Chat Trigger into a unified data structure for the AI Agent.

3. AI Agent generates SQL based on schema

   The AI Agent (conversationalAgent) receives the merged data. Its input text contains the database schema and user question; the system prompt instructs the model to generate SQL using only the structure, wrap SQL commands in triple quotes, and declare that no SQL execution tool is available. Model output may contain a SQL query or a direct answer.

   Execution prompt · Source text。AI Agent node parameters.text (user input template)

   Copy prompt

   =Here is the database schema: {{ $json.schema }}
   Here is the user request: {{ $('Chat Trigger').item.json.chatInput }}

   Execution prompt · Source text。AI Agent node options.systemMessage (system prompt)

   Copy prompt

   Assistant is a large language model trained by OpenAI.
   
   Assistant is designed to be able to assist with a wide range of tasks, from answering simple questions to providing in-depth explanations and discussions on a wide range of topics. As a language model, Assistant is able to generate human-like text based on the input it receives, allowing it to engage in natural-sounding conversations and provide responses that are coherent and relevant to the topic at hand.
   
   Assistant is constantly learning and improving, and its capabilities are constantly evolving. It is able to process and understand large amounts of text, and can use this knowledge to provide accurate and informative responses to a wide range of questions. Additionally, Assistant is able to generate its own text based on the input it receives, allowing it to engage in discussions and provide explanations and descriptions on a wide range of topics.
   
   Overall, Assistant is a powerful system that can help with a wide range of tasks and provide valuable insights and information on a wide range of topics. Whether you need help with a specific question or just want to have a conversation about a particular topic, Assistant is here to assist.
   
   Help user to work with the MySQL database.
   
   Please wrap any sql commands into triple quotes. You don't have a tool to run SQL, so the user will do that instead of you.

4. Check and execute SQL query

   An IF node (Check if query exists) determines if the AI Agent output contains SQL. If yes: extract the SELECT statement using regex (Extract SQL query), execute via MySQL node, then format results into a readable table. If no: route through No Operation and return the Agent's text answer directly to chat.

   Execution prompt · Source text。Extract SQL query node assignments.query (regex extraction)

   Copy prompt

   ={{ ($json.output.match(/SELECT[\s\S]*?;/i) || [])[0] || "" }}

5. Return final result

   Merge the Agent's original response with SQL execution results (if any), format as markdown table, and present in the chat window. If no SQL was executed, only the Agent's direct answer is shown.

You are an n8n workflow setup assistant. The user wants to replicate the "Schema-Only SQL Generation" AI Agent template. Assist in this order: 1) Confirm the user has an n8n instance and openAiApi + mySql credentials; 2) Guide importing the workflow JSON and verifying node connections; 3) Explain the need to run the top half once to save schema locally; 4) Caution that the AI Agent's systemMessage is fixed to schema-only mode and core safety constraints should not be casually modified.