This workflow follows the Agent → Chat Trigger recipe pattern — see all workflows that pair these two integrations.
The workflow JSON
Copy or download the full n8n JSON below. Paste it into a new n8n workflow, add your credentials, activate. Full import guide →
{
"name": "updated!",
"nodes": [
{
"parameters": {
"content": "## User Query Workflow\n**User types message -> get metadata -> generate custom prompt with metadata -> generate sql -> get results",
"height": 176,
"width": 320,
"color": 5
},
"type": "n8n-nodes-base.stickyNote",
"position": [
4528,
-432
],
"typeVersion": 1,
"id": "e7508b83-c4e7-476c-9699-0ad73e563e7f",
"name": "Sticky Note"
},
{
"parameters": {
"modelName": "models/gemini-2.0-flash-lite-preview",
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.lmChatGoogleGemini",
"typeVersion": 1,
"position": [
2680,
-1008
],
"id": "ec6f7e14-ce19-4f06-af0f-bdc703abf1d7",
"name": "Google Gemini Chat Model",
"credentials": {
"googlePalmApi": {
"name": "<your credential>"
}
}
},
{
"parameters": {
"modelName": "models/gemini-2.5-flash-lite",
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.lmChatGoogleGemini",
"typeVersion": 1,
"position": [
4216,
-1008
],
"id": "ab03816b-30c6-4b00-a308-88fe35f20084",
"name": "Google Gemini Chat Model1",
"credentials": {
"googlePalmApi": {
"name": "<your credential>"
}
}
},
{
"parameters": {
"promptType": "define",
"text": "={{ $json.prompt }}",
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.agent",
"typeVersion": 2,
"position": [
2672,
-1232
],
"id": "a0651658-4450-4c7e-b34d-0301bc0b11b5",
"name": "AI Agent2"
},
{
"parameters": {
"promptType": "define",
"text": "=You are a helpful assistant that transforms SQL query results into clear, professional summaries for end users. \n\nInput: \nobjects \n{{ Columns = ($json[\"rows\"] && $json[\"rows\"].length > 0) ? $json[\"rows\"].last().keys() : [] }} \n{{ Values = ($json[\"rows\"] && $json[\"rows\"].length > 0) ? $json[\"rows\"].map((item) => item.values()) : [] }} \n\nInstructions: \n\n\u26a0\ufe0f Important: \n- If `Values` is an **empty array**, it means the SQL query executed successfully but **returned no rows**. \n- Do **not** ask the user to provide data again. Instead, handle it gracefully as a \u201cno results\u201d case. \n\n### \ud83d\udeab If No Rows Are Returned (`Values` is empty): \n1. Respond politely and formally. \n2. Provide a helpful message, such as: \n - \"No matching records were found.\" \n - \"Are you sure the values exist in the database?\" \n - \"Perhaps you meant a different filter or table?\" \n3. Suggest possible next steps (e.g., check spelling, broaden filters, verify data availability). \n4. Ensure the response is nicely formatted with line spacing and bullet points. \n5. Close with a supportive sentence, e.g.: \n - \"Please try refining your query and run it again.\" \n\n### \u2705 If Rows Are Returned: \nGenerate a well-structured summary with the following components: \n\n1. \ud83d\udccb Summary Explanation \n - Provide a concise explanation of what the returned data represents. \n\n2. \ud83e\udde0 Analytics \n - Analyze the data to identify patterns, trends, or notable points. \n\n3. \ud83d\udcca Formatted Results Table \n - Display the raw SQL results in a clean, readable Markdown table. \n - Ensure: \n * Proper column headers (converted to Title Case) \n * Neat alignment \n * No truncation or overflow \n * Added padding & extra spacing between columns \n * Spacing after data rows \n\nFinal Note: Ensure the entire response is well-formatted, user-friendly, and suitable for presentation or reporting purposes. \n",
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.agent",
"typeVersion": 2,
"position": [
4144,
-1232
],
"id": "9a085462-f53c-4829-b82b-1ab3701904fc",
"name": "AI Agent"
},
{
"parameters": {
"sessionIdType": "customKey",
"sessionKey": "="
},
"type": "@n8n/n8n-nodes-langchain.memoryBufferWindow",
"typeVersion": 1.3,
"position": [
2808,
-1008
],
"id": "de532b02-ed78-486c-a1fa-e0a9c8079999",
"name": "Simple Memory",
"disabled": true
},
{
"parameters": {
"jsCode": "return [\n {\n json: {\n query: $input.first().json.body.message\n }\n }\n];\n"
},
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
2000,
-1128
],
"id": "4b5eb204-e90c-4e87-b645-f419042e7dac",
"name": "Convert to JSON"
},
{
"parameters": {
"jsCode": "const userQuery = $('Convert to JSON').first().json.query;\nconst metadata = $input.first().json.context;\n\nconst prompt = `\nYou are a helpful assistant. Your job is to translate natural language questions into SQL queries that can be executed on an Oracle database.\n\nUser request: ${userQuery}\n\nRelevant DB Metadata (tables, columns, FKs):\n${JSON.stringify(metadata, null, 2)}\n\nYour task is to output only a valid SQL query using Oracle syntax. \nDo not include any explanation, comments, or text outside the query.\n\nWhen generating queries:\n* Always respect foreign key relationships and join related tables when needed to retrieve meaningful or complete information.\n* If a column is a foreign key, consider joining with the referenced table to get descriptive fields (e.g., product name instead of product_id).\n* Use appropriate table aliases for readability if joining multiple tables.\n* Ensure the query reflects the intent of the user's question, even if it requires joining across multiple related tables.\n* Ensure not to include ID fields unless specifically asked for.\n\nAlways output only a raw SQL query. Do not include any markdown formatting, comments, or extra explanation \u2014 just the query.\n`;\n\nreturn [{ json: { prompt } }];\n"
},
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
2448,
-1128
],
"id": "23850205-19a5-4ac4-beda-e427ef44354a",
"name": "Generate Custom Prompt"
},
{
"parameters": {
"jsCode": "const items = $input.all();\n\nconst updatedItems = items.map((item) => {\n let query = item?.json?.output;\n if (!query) return item;\n\n // Remove triple backticks and language tag\n query = query\n .replace(/^```[\\w]*\\r?\\n/i, \"\")\n .replace(/\\r?\\n```$/, \"\");\n\n // Convert escaped \\n to real newlines\n query = query.replace(/\\\\n/g, \"\\n\").trim();\n\n // Remove trailing semicolons\n query = query.replace(/;+\\s*$/, \"\").trim();\n\n // Ensure query starts with SELECT\n if (!/^SELECT\\s/i.test(query)) {\n throw new Error(`Invalid query format: must start with SELECT`);\n }\n\n // Case-insensitive matching: transform \"col = 'value'\" into \"UPPER(col) = UPPER('value')\"\n query = query.replace(\n /(\\bWHERE\\b[\\s\\S]*?)(\\b\\w+\\b)\\s*=\\s*'([^']+)'/i,\n (match, wherePart, column, value) => {\n return `${wherePart}UPPER(${column}) = UPPER('${value}')`;\n }\n );\n\n // Normalize spaces\n item.json.output = query.replace(/\\s+/g, \" \").trim();\n\n return item;\n});\n\nreturn updatedItems;\n"
},
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
3024,
-1128
],
"id": "1b6fcde4-4801-4fc7-a712-c42d6c52c005",
"name": "Clean Up SQL"
},
{
"parameters": {
"method": "POST",
"url": "http://127.0.0.1:8000/similar-metadata",
"sendBody": true,
"bodyParameters": {
"parameters": [
{
"name": "query",
"value": "={{ $json.query }}"
}
]
},
"options": {
"timeout": 80000
}
},
"type": "n8n-nodes-base.httpRequest",
"typeVersion": 4.2,
"position": [
2224,
-1128
],
"id": "8d421544-e097-4675-ad08-42d97b4c17a0",
"name": "Get Table Metadata"
},
{
"parameters": {
"url": "http://127.0.0.1:8000/db-direct?",
"sendQuery": true,
"queryParameters": {
"parameters": [
{
"name": "query",
"value": "={{$json.output}}"
}
]
},
"sendHeaders": true,
"headerParameters": {
"parameters": [
{
"name": "Authorization",
"value": "={{ $('Webhook').item.json.body.headers.Authorization }}"
}
]
},
"options": {}
},
"type": "n8n-nodes-base.httpRequest",
"typeVersion": 4.2,
"position": [
3248,
-1128
],
"id": "b1658851-3437-44c3-a6e8-49924c13ad2f",
"name": "Get Output from Oracle"
},
{
"parameters": {
"jsCode": "const items = $input.all();\n\n\n\n// Flatten each row to a readable format\n// const rows = temp.map((item, index) => {\n// const row = item;\n// const readable = Object.entries(row)\n// .map(([key, value]) => `**${key}**: ${value}`)\n// .join(', ');\n// return `Row ${index + 1}: ${readable}`;\n// });\n\nreturn [\n {\n json: {\n success: items[0].json.success,\n rows: items[0].json.results\n }\n }\n];\n"
},
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
3472,
-1128
],
"id": "64bfa850-17f4-45a4-bcf6-316088ddb4d1",
"name": "Get Outputs"
},
{
"parameters": {
"conditions": {
"options": {
"caseSensitive": true,
"leftValue": "",
"typeValidation": "loose",
"version": 2
},
"conditions": [
{
"id": "7e1204dc-a1b6-4c29-acfa-c64fb2a5c01b",
"leftValue": "={{ $json.success }}",
"rightValue": "",
"operator": {
"type": "boolean",
"operation": "true",
"singleValue": true
}
}
],
"combinator": "and"
},
"looseTypeValidation": true,
"options": {}
},
"type": "n8n-nodes-base.if",
"typeVersion": 2.2,
"position": [
3696,
-1128
],
"id": "549ac108-0d59-482d-816a-a211a9db0ece",
"name": "Success Test"
},
{
"parameters": {
"conditions": {
"options": {
"caseSensitive": true,
"leftValue": "",
"typeValidation": "loose",
"version": 2
},
"conditions": [
{
"id": "9f40634d-dafb-4ac7-9af3-d15c45d21897",
"leftValue": "={{ $json.rows.error }}",
"rightValue": "",
"operator": {
"type": "boolean",
"operation": "notExists",
"singleValue": true
}
}
],
"combinator": "and"
},
"looseTypeValidation": true,
"options": {}
},
"type": "n8n-nodes-base.if",
"typeVersion": 2.2,
"position": [
3920,
-1128
],
"id": "a8653e5a-70cd-47f6-830a-5eade09a93f7",
"name": "If"
},
{
"parameters": {
"respondWith": "allIncomingItems",
"options": {
"responseCode": 200,
"responseHeaders": {
"entries": [
{
"name": "content-type",
"value": "text"
}
]
}
}
},
"type": "n8n-nodes-base.respondToWebhook",
"typeVersion": 1.4,
"position": [
4496,
-1128
],
"id": "5c4bbddb-2395-46dc-ab1a-187942c93d91",
"name": "Respond to Webhook"
},
{
"parameters": {
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.chatTrigger",
"typeVersion": 1.1,
"position": [
1776,
-696
],
"id": "e9605e6e-c4b6-40cf-b610-d3ee84f13a23",
"name": "When chat message received"
},
{
"parameters": {
"modelName": "models/gemini-2.0-flash-lite-preview",
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.lmChatGoogleGemini",
"typeVersion": 1,
"position": [
2680,
-472
],
"id": "c6626c41-8d76-4bc9-9edc-fb995fac2b91",
"name": "Google Gemini Chat Model2",
"credentials": {
"googlePalmApi": {
"name": "<your credential>"
}
}
},
{
"parameters": {
"modelName": "models/gemini-2.5-flash-lite",
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.lmChatGoogleGemini",
"typeVersion": 1,
"position": [
4216,
-472
],
"id": "b9aaa08d-7db1-44d7-b5d2-d567a18fab68",
"name": "Google Gemini Chat Model3",
"credentials": {
"googlePalmApi": {
"name": "<your credential>"
}
}
},
{
"parameters": {
"promptType": "define",
"text": "={{ $json.prompt }}",
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.agent",
"typeVersion": 2,
"position": [
2672,
-696
],
"id": "03cd1937-08ea-40a3-b6bc-8d25b47fc27d",
"name": "AI Agent3"
},
{
"parameters": {
"promptType": "define",
"text": "=You are a helpful assistant that transforms SQL query results into clear, professional summaries for end users. Given the following input:\nobjects\n{{ Columns = $json[\"rows\"].last().keys()}}\n{{ Values = $json[\"rows\"].map((item)=> item.values()); \n \n}}\n\nGenerate a well-structured summary with the following components: 1. \ud83d\udccb **Summary Explanation** Provide a concise explanation of what the returned data represents. \n\n2. \ud83e\udde0 **Analytics** Provide analytics on the datato find patterns.\n\n3. \ud83d\udcca **Formatted Results Table** Display the raw SQL results in a clean and readable Markdown table, ensuring: \n* Proper column headers \n* Neat alignment \n* No truncation or overflow \n* Added Padding and Extra Spacing between columns !IMPORTANT \n* Add spacing after data\n* Convert column headers to Title Case Ensure the entire response is well-formatted, user-friendly, and suitable for presentation or reporting purposes. The output is intended for markdown",
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.agent",
"typeVersion": 2,
"position": [
4144,
-696
],
"id": "80bbf0be-6d3e-46d2-a687-2906cbe8c9c4",
"name": "AI Agent1"
},
{
"parameters": {
"conditions": {
"options": {
"caseSensitive": true,
"leftValue": "",
"typeValidation": "strict",
"version": 2
},
"conditions": [
{
"id": "90a9f433-8819-4b1c-97b8-2be0212515c4",
"leftValue": "={{ $json.success }}",
"rightValue": "",
"operator": {
"type": "boolean",
"operation": "true",
"singleValue": true
}
}
],
"combinator": "and"
},
"options": {}
},
"type": "n8n-nodes-base.if",
"typeVersion": 2.2,
"position": [
2448,
-1456
],
"id": "efa8b497-04fb-4521-83c4-494754d7c396",
"name": "If3"
},
{
"parameters": {
"url": "http://127.0.0.1:8000/refresh-metadata",
"options": {
"timeout": 80000
}
},
"type": "n8n-nodes-base.httpRequest",
"typeVersion": 4.2,
"position": [
2000,
-1456
],
"id": "ca80629d-6b7d-4ae6-8606-2fe377c731e2",
"name": "Refresh DB Metadata1"
},
{
"parameters": {
"method": "POST",
"url": "http://127.0.0.1:8000/embed-metadata?owner=CHATBOT_USER",
"options": {
"timeout": 800000
}
},
"type": "n8n-nodes-base.httpRequest",
"typeVersion": 4.2,
"position": [
2224,
-1456
],
"id": "df136855-2e88-48b8-ab6a-bda17cc1eba3",
"name": "Run Embedding Pipeline1"
},
{
"parameters": {
"method": "=POST",
"url": "http://127.0.0.1:8000/similar-metadata",
"sendBody": true,
"specifyBody": "json",
"jsonBody": "{\n \"query\": \"Vehicles\"\n}\n",
"options": {}
},
"type": "n8n-nodes-base.httpRequest",
"typeVersion": 4.2,
"position": [
2672,
-1456
],
"id": "d38f29a4-89bf-430a-9f74-5beabf04a625",
"name": "Test Similar Metadata1"
},
{
"parameters": {
"content": "## Metadata Ingestion Workflow:\nAdds Embeddings for Metadata and Tests with one table name.\nTo Be Run: Periodically or On Change in Tables.",
"height": 120,
"width": 380,
"color": 3
},
"type": "n8n-nodes-base.stickyNote",
"position": [
3856,
-1456
],
"typeVersion": 1,
"id": "cfe83e37-59d1-4051-8b53-6885dbadda62",
"name": "Metadata Ingestion1"
},
{
"parameters": {
"sessionIdType": "customKey",
"sessionKey": "={{ $('When chat message received').item.json.sessionId }}"
},
"type": "@n8n/n8n-nodes-langchain.memoryBufferWindow",
"typeVersion": 1.3,
"position": [
2808,
-472
],
"id": "fac085bc-b5f3-4547-a0c8-05a61d702c9a",
"name": "Simple Memory1"
},
{
"parameters": {
"jsCode": "return [\n {\n json: {\n query: $input.first().json.chatInput\n }\n }\n];\n"
},
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
2000,
-696
],
"id": "b92ce305-0ed1-49c3-9b62-41d7052d5773",
"name": "Convert to JSON1"
},
{
"parameters": {
"jsCode": "const userQuery = $('Convert to JSON1').first().json.query;\nconst metadata = $input.first().json.context;\n\nconst prompt = `\nYou are a helpful assistant. Your job is to translate natural language questions into SQL queries that can be executed on an Oracle database.\n\nUser request: ${userQuery}\n\nRelevant DB Metadata (tables, columns, FKs):\n${JSON.stringify(metadata, null, 2)}\n\nYour task is to output only a valid SQL query using Oracle syntax. \nDo not include any explanation, comments, or text outside the query.\n\nWhen generating queries:\n* Always respect foreign key relationships and join related tables when needed to retrieve meaningful or complete information.\n* If a column is a foreign key, consider joining with the referenced table to get descriptive fields (e.g., product name instead of product_id).\n* Use appropriate table aliases for readability if joining multiple tables.\n* Ensure the query reflects the intent of the user's question, even if it requires joining across multiple related tables.\n* Ensure not to include ID fields unless specifically asked for.\n\nAlways output only a raw SQL query. Do not include any markdown formatting, comments, or extra explanation \u2014 just the query.\n`;\n\nreturn [{ json: { prompt } }];\n"
},
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
2448,
-696
],
"id": "8989b430-578e-4286-9f72-c9df06dd07a9",
"name": "Generate Custom Prompt1"
},
{
"parameters": {
"jsCode": "const items = $input.all();\n\nconst updatedItems = items.map((item) => {\n let query = item?.json?.output;\n\n if (!query) return item;\n\n // Clean up the query first\n query = query.replace(/^```sql\\\\n|^```sql\\n|^```sql\\r\\n/, \"\")\n .replace(/\\\\n```$|\\n```$|\\r\\n```$/, \"\")\n .replace(/\\\\n/g, \"\\n\")\n .replace(/;\\s*$/, \"\")\n .trim();\n\n // Convert to uppercase to check if it's a SELECT query\n const upperQuery = query.toUpperCase();\n \n if (upperQuery.startsWith(\"SELECT\")) {\n // Handle case sensitivity for table and column names\n query = query.replace(/\\b(FROM|JOIN)\\s+(\\w+)/gi, (match, keyword, tableName) => {\n return `${keyword} \"${tableName.toUpperCase()}\"`;\n });\n\n // Handle column names in SELECT clause\n query = query.replace(/\\bSELECT\\s+(.*?)\\bFROM/gi, (match, columnsPart) => {\n const processedColumns = columnsPart.split(',').map(col => {\n const trimmedCol = col.trim();\n if (trimmedCol && !trimmedCol.includes('(') && !trimmedCol.includes('*')) {\n return `\"${trimmedCol.toUpperCase()}\"`;\n }\n return trimmedCol;\n }).join(', ');\n \n return `SELECT ${processedColumns} FROM`;\n });\n\n // Handle WHERE clause with string comparisons - make them case-insensitive\n if (query.includes(\"WHERE\")) {\n query = query.replace(/\\bWHERE\\s+(.*)/gi, (match, whereClause) => {\n // Handle = comparisons with strings\n let processedWhere = whereClause.replace(/(\\w+)\\s*=\\s*'([^']*)'/gi, (conditionMatch, column, value) => {\n if (isNaN(value) && value !== 'NULL' && value !== 'null') {\n return `UPPER(\"${column.toUpperCase()}\") = UPPER('${value}')`;\n }\n return conditionMatch;\n });\n\n // Handle LIKE comparisons\n processedWhere = processedWhere.replace(/(\\w+)\\s+LIKE\\s+'([^']*)'/gi, (conditionMatch, column, value) => {\n return `UPPER(\"${column.toUpperCase()}\") LIKE UPPER('${value}')`;\n });\n\n return `WHERE ${processedWhere}`;\n });\n }\n }\n\n // Clean up spacing\n query = query.replace(/\\s+/g, \" \").trim();\n \n item.json.output = query;\n return item;\n});\n\nreturn updatedItems;"
},
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
3024,
-696
],
"id": "bd8bae3a-1c11-4373-954a-23cfce9fdeb5",
"name": "Clean Up SQL1"
},
{
"parameters": {
"method": "POST",
"url": "http://127.0.0.1:8000/similar-metadata",
"sendBody": true,
"bodyParameters": {
"parameters": [
{
"name": "query",
"value": "={{ $json.query }}"
}
]
},
"options": {}
},
"type": "n8n-nodes-base.httpRequest",
"typeVersion": 4.2,
"position": [
2224,
-696
],
"id": "c8f0ead1-df0b-4f56-91e2-c53f4589f82c",
"name": "Get Table Metadata1"
},
{
"parameters": {
"url": "http://127.0.0.1:8000/db-direct?",
"sendQuery": true,
"queryParameters": {
"parameters": [
{
"name": "query",
"value": "={{$json.output}}"
}
]
},
"options": {}
},
"type": "n8n-nodes-base.httpRequest",
"typeVersion": 4.2,
"position": [
3248,
-696
],
"id": "59d193e8-bc74-4e9c-a21c-c67d24855184",
"name": "Get Output from Oracle1"
},
{
"parameters": {
"jsCode": "const items = $input.all();\n\n\n\n// Flatten each row to a readable format\n// const rows = temp.map((item, index) => {\n// const row = item;\n// const readable = Object.entries(row)\n// .map(([key, value]) => `**${key}**: ${value}`)\n// .join(', ');\n// return `Row ${index + 1}: ${readable}`;\n// });\n\nreturn [\n {\n json: {\n success: items[0].json.success,\n rows: items[0].json.results\n }\n }\n];\n"
},
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
3472,
-696
],
"id": "2f1c1a40-df49-4c3c-8eac-cad477f9e2b5",
"name": "Get Outputs1"
},
{
"parameters": {
"conditions": {
"options": {
"caseSensitive": true,
"leftValue": "",
"typeValidation": "loose",
"version": 2
},
"conditions": [
{
"id": "7e1204dc-a1b6-4c29-acfa-c64fb2a5c01b",
"leftValue": "={{ $json.success }}",
"rightValue": "",
"operator": {
"type": "boolean",
"operation": "true",
"singleValue": true
}
}
],
"combinator": "and"
},
"looseTypeValidation": true,
"options": {}
},
"type": "n8n-nodes-base.if",
"typeVersion": 2.2,
"position": [
3696,
-696
],
"id": "b32b212a-3c8a-4116-b7c5-f2890cfbd94d",
"name": "Success Test1"
},
{
"parameters": {
"conditions": {
"options": {
"caseSensitive": true,
"leftValue": "",
"typeValidation": "loose",
"version": 2
},
"conditions": [
{
"id": "9f40634d-dafb-4ac7-9af3-d15c45d21897",
"leftValue": "={{ $json.rows.error }}",
"rightValue": "",
"operator": {
"type": "string",
"operation": "notExists",
"singleValue": true
}
}
],
"combinator": "and"
},
"looseTypeValidation": true,
"options": {}
},
"type": "n8n-nodes-base.if",
"typeVersion": 2.2,
"position": [
3920,
-696
],
"id": "aaac553f-4619-4bbb-aea6-38cd6b5f66ee",
"name": "If1"
},
{
"parameters": {
"content": "## Test Query Workflow\nSee Responses and accuracy using n8n chat only. ",
"height": 220,
"width": 220
},
"type": "n8n-nodes-base.stickyNote",
"position": [
5056,
-128
],
"typeVersion": 1,
"id": "3d1af3a7-66c1-4f86-b562-3bf7b44dbfeb",
"name": "Sticky Note3"
},
{
"parameters": {},
"type": "n8n-nodes-base.manualTrigger",
"typeVersion": 1,
"position": [
1776,
-1456
],
"id": "bcda7146-af80-4013-a445-e6e08ef1fe22",
"name": "When clicking \u2018Execute workflow\u2019"
},
{
"parameters": {
"content": "## Webhook Query Workflow\n**Gets response from frontend -> activates -> get metadata -> generate custom prompt with metadata -> generate sql -> get results",
"width": 320
},
"type": "n8n-nodes-base.stickyNote",
"position": [
1424,
-928
],
"typeVersion": 1,
"id": "14b31d8c-8ee0-4f75-93ee-4f837d5b0634",
"name": "Sticky Note2"
},
{
"parameters": {
"httpMethod": "POST",
"path": "03e650c2-18be-4c37-903a-4e99bddcc8b1",
"responseMode": "responseNode",
"options": {}
},
"type": "n8n-nodes-base.webhook",
"typeVersion": 2.1,
"position": [
1792,
-1104
],
"id": "f7a64be5-a40b-4e7e-902e-bd8731796bd8",
"name": "Webhook",
"alwaysOutputData": false
},
{
"parameters": {
"content": "## User Query Workflow\n**User types message -> get metadata -> generate custom prompt with metadata -> generate sql -> get results",
"width": 320,
"color": 5
},
"type": "n8n-nodes-base.stickyNote",
"position": [
2592,
1072
],
"typeVersion": 1,
"id": "a7a91390-640a-4d77-8933-9c1b10f8e844",
"name": "Sticky Note4"
},
{
"parameters": {
"content": "## Improvements \n1. Accuracy of similar-metadata function \n2. Prompts for final output agent"
},
"type": "n8n-nodes-base.stickyNote",
"position": [
6016,
1536
],
"typeVersion": 1,
"id": "ce9607fb-23c3-4ddf-a2e2-98703ea636a1",
"name": "Sticky Note5"
}
],
"connections": {
"Google Gemini Chat Model": {
"ai_languageModel": [
[
{
"node": "AI Agent2",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"Google Gemini Chat Model1": {
"ai_languageModel": [
[
{
"node": "AI Agent",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"AI Agent2": {
"main": [
[
{
"node": "Clean Up SQL",
"type": "main",
"index": 0
}
]
]
},
"Simple Memory": {
"ai_memory": [
[
{
"node": "AI Agent2",
"type": "ai_memory",
"index": 0
}
]
]
},
"Convert to JSON": {
"main": [
[
{
"node": "Get Table Metadata",
"type": "main",
"index": 0
}
]
]
},
"Generate Custom Prompt": {
"main": [
[
{
"node": "AI Agent2",
"type": "main",
"index": 0
}
]
]
},
"Clean Up SQL": {
"main": [
[
{
"node": "Get Output from Oracle",
"type": "main",
"index": 0
}
]
]
},
"Get Table Metadata": {
"main": [
[
{
"node": "Generate Custom Prompt",
"type": "main",
"index": 0
}
]
]
},
"Get Output from Oracle": {
"main": [
[
{
"node": "Get Outputs",
"type": "main",
"index": 0
}
]
]
},
"Get Outputs": {
"main": [
[
{
"node": "Success Test",
"type": "main",
"index": 0
}
]
]
},
"Success Test": {
"main": [
[
{
"node": "If",
"type": "main",
"index": 0
}
]
]
},
"If": {
"main": [
[
{
"node": "AI Agent",
"type": "main",
"index": 0
}
]
]
},
"AI Agent": {
"main": [
[
{
"node": "Respond to Webhook",
"type": "main",
"index": 0
}
]
]
},
"When chat message received": {
"main": [
[
{
"node": "Convert to JSON1",
"type": "main",
"index": 0
}
]
]
},
"Google Gemini Chat Model2": {
"ai_languageModel": [
[
{
"node": "AI Agent3",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"Google Gemini Chat Model3": {
"ai_languageModel": [
[
{
"node": "AI Agent1",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"AI Agent3": {
"main": [
[
{
"node": "Clean Up SQL1",
"type": "main",
"index": 0
}
]
]
},
"If3": {
"main": [
[
{
"node": "Test Similar Metadata1",
"type": "main",
"index": 0
}
]
]
},
"Refresh DB Metadata1": {
"main": [
[
{
"node": "Run Embedding Pipeline1",
"type": "main",
"index": 0
}
]
]
},
"Run Embedding Pipeline1": {
"main": [
[
{
"node": "If3",
"type": "main",
"index": 0
}
]
]
},
"Simple Memory1": {
"ai_memory": [
[
{
"node": "AI Agent3",
"type": "ai_memory",
"index": 0
}
]
]
},
"Convert to JSON1": {
"main": [
[
{
"node": "Get Table Metadata1",
"type": "main",
"index": 0
}
]
]
},
"Generate Custom Prompt1": {
"main": [
[
{
"node": "AI Agent3",
"type": "main",
"index": 0
}
]
]
},
"Clean Up SQL1": {
"main": [
[
{
"node": "Get Output from Oracle1",
"type": "main",
"index": 0
}
]
]
},
"Get Table Metadata1": {
"main": [
[
{
"node": "Generate Custom Prompt1",
"type": "main",
"index": 0
}
]
]
},
"Get Output from Oracle1": {
"main": [
[
{
"node": "Get Outputs1",
"type": "main",
"index": 0
}
]
]
},
"Get Outputs1": {
"main": [
[
{
"node": "Success Test1",
"type": "main",
"index": 0
}
]
]
},
"Success Test1": {
"main": [
[
{
"node": "If1",
"type": "main",
"index": 0
}
]
]
},
"If1": {
"main": [
[
{
"node": "AI Agent1",
"type": "main",
"index": 0
}
]
]
},
"When clicking \u2018Execute workflow\u2019": {
"main": [
[
{
"node": "Refresh DB Metadata1",
"type": "main",
"index": 0
}
]
]
},
"Webhook": {
"main": [
[
{
"node": "Convert to JSON",
"type": "main",
"index": 0
}
]
]
}
},
"active": true,
"settings": {
"executionOrder": "v1"
},
"versionId": "37aa27c5-80ae-435f-97cd-74ef396a2b18",
"meta": {
"templateCredsSetupCompleted": true
},
"id": "VdEhXZSWb3M6wi7Q",
"tags": []
}
Credentials you'll need
Each integration node will prompt for credentials when you import. We strip credential IDs before publishing — you'll add your own.
googlePalmApi
Pro
For the full experience including quality scoring and batch install features for each workflow upgrade to Pro
About this workflow
updated!. Uses lmChatGoogleGemini, agent, memoryBufferWindow, httpRequest. Webhook trigger; 40 nodes.
Source: https://github.com/RafayRandhawa/AI-Oracle-Chatbot/blob/9fceeceba174158581c8e84ce9d77db1f62d5e27/workflows/updated!.json — original creator credit. Request a take-down →
Related workflows
Workflows that share integrations, category, or trigger type with this one. All free to copy and import.
veo limpo new. Uses moveBinaryData, httpRequest, chatTrigger, baserow. Webhook trigger; 36 nodes.
Move Binary Data, HTTP Request, Chat Trigger +8
N8N-Chat. Uses jwt, agent, memoryBufferWindow, lmChatGoogleGemini. Webhook trigger; 66 nodes.
Jwt, Agent, Memory Buffer Window +2
Catat Keuangan Keluarga. Uses chatTrigger, agent, lmChatGoogleGemini, outputParserStructured. Webhook trigger; 55 nodes.
Chat Trigger, Agent, Google Gemini Chat +5
This workflow acts as an AI-powered research assistant that takes a topic from the user, performs multi-step intelligent research, and stores the final report in Notion. It uses advanced search, conte
Memory Buffer Window, Output Parser Structured, Agent +6
This workflow transforms WhatsApp into a powerful personal AI using n8n + Green-API. Send text or voice messages — the assistant understands intent and handles daily tasks automatically. 💰 Expense & i
Tool Calculator, Google Sheets Tool, OpenAI Chat +10