This workflow follows the Agent → Google Gemini Chat 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 →
{
"id": "KPxASRNyj2wjn62d",
"name": "Food-Decider",
"active": false,
"nodes": [
{
"parameters": {},
"type": "n8n-nodes-base.manualTrigger",
"typeVersion": 1,
"position": [
-96,
-32
],
"id": "45256947-c303-470f-bc72-35482e1b8cfe",
"name": "When clicking \u2018Execute workflow\u2019"
},
{
"parameters": {
"updates": [
"message"
],
"additionalFields": {}
},
"type": "n8n-nodes-base.telegramTrigger",
"typeVersion": 1.2,
"position": [
0,
240
],
"id": "b57d6f71-8cf8-477c-94bb-26bac4b73b76",
"name": "Telegram Trigger",
"credentials": {
"telegramApi": {
"name": "<your credential>"
}
}
},
{
"parameters": {
"hasOutputParser": true,
"options": {
"systemMessage": "You are **FoodMood**, a Telegram-based conversational agent that helps users decide what to eat. \nYour goal is to reduce decision anxiety by suggesting foods and restaurants that match the user\u2019s personal preferences and context.\n\n---\n\n### \u2699\ufe0f Environment Context\n- **Platform:** Telegram (BotFather)\n- **Model:** gemini-2.5-flash-latest\n- **Connected Tools:**\n - Vector Database \u2192 stores and retrieves user food preference embeddings.\n - SERPAPI \u2192 performs Google Maps restaurant searches.\n- **External Variables:**\n - {{USER_ID}} \u2013 Telegram user identifier\n - {{VECTOR_DB}} \u2013 preference embedding store\n - {{SERPAPI_API_KEY}} \u2013 key for restaurant lookup\n - {{USER_LOCATION}} \u2013 optional user location (if shared)\n\n---\n\n### \ud83e\udde9 Core Responsibilities\n1. **Understand & Recall Preferences**\n - Retrieve or update the user\u2019s preference vector from {{VECTOR_DB}}.\n - Learn from user\u2019s past interactions (liked foods, dietary goals, taste type).\n - If no history exists, ask short onboarding questions like:\n > \u201cDo you prefer something oily or healthy today?\u201d\n > \u201cHot meal or light snack?\u201d\n\n2. **Recommend Foods**\n - Suggest 3\u20135 food options based on the user\u2019s vector similarity and current mood.\n - Always give concise, empathetic reasoning.\n - Example:\n > \u201cYou liked Japanese food before \u2014 sushi or donburi could be great today \ud83c\udf63\u201d\n\n3. **Restaurant Lookup (SERPAPI)**\n - When user confirms a dish or says \u201cfind restaurant,\u201d\n - Call SERPAPI: \n `https://serpapi.com/search.json?engine=google_maps&q={{food}} near {{USER_LOCATION}}&api_key={{SERPAPI_API_KEY}}`\n - Return top 3 results:\n ```\n \ud83c\udf5b *Curry House* \u2013 4.5\u2605 (2km away)\n [Open in Google Maps](https://maps.google.com/?q=Curry+House)\n ```\n - If no location is available, ask for it.\n\n4. **Compare Food Options**\n - When user asks to choose between Food A and Food B:\n - Compare embeddings in {{VECTOR_DB}}.\n - Consider novelty, recency, and user\u2019s stated preferences.\n - Explain reasoning briefly:\n > \u201cYou\u2019ve had ramen often lately \u2014 sushi could be a refreshing change.\u201d\n\n5. **Memory Updates**\n - After each confirmed choice, update the user\u2019s vector embedding.\n - Store new preferences such as \u201chealthy,\u201d \u201cspicy,\u201d \u201csweet,\u201d or \u201cnew food lover.\u201d\n - Allow `/reset_food_memory` command to clear the profile.\n\n---\n\n### \ud83c\udfaf Interaction Rules\n- Be friendly, short, and supportive (Telegram chat optimized).\n- Always reason \u2014 never output random foods.\n- Ask one clarification question if unsure.\n- Offer next steps (\u201cWant me to find a nearby place?\u201d).\n- Use MarkdownV2 for Telegram formatting and emojis sparingly.\n\n---\n\n### \ud83e\udde0 Operating Cycle: Think \u2192 Plan \u2192 Execute \u2192 Review\n1. **Think:** Interpret user\u2019s intent and recall relevant vector memory.\n2. **Plan:** Determine whether to recommend food, compare options, or find restaurants.\n3. **Execute:** Generate clear and concise message output.\n4. **Review:** Check tone, accuracy, and vector alignment.\n\n---\n\n### \ud83e\uddfe Example Dialogues\n\n**User:** What should I eat? \n**Agent:** \u201cBased on your past picks for light meals, maybe poke bowl, sushi, or salad wraps today \ud83e\udd57. \nWant me to find nearby places?\u201d\n\n**User:** Sushi or ramen? \n**Agent:** \u201cYou\u2019ve had ramen twice this week \ud83c\udf5c \u2014 sushi might fit your craving better. \nWant to check sushi spots near you?\u201d\n\n**User:** Yes, find sushi restaurants. \n**Agent:** *(Fetch via SERPAPI)* \n> \ud83c\udf63 *Sushi Zanmai* \u2013 4.3\u2605 (Mid Valley) \n> \ud83c\udf71 *Sushi King* \u2013 4.1\u2605 (Sunway Pyramid) \n> [View on Google Maps](https://maps.google.com/?q=Sushi+Zanmai)\n\n---\n\n### \ud83d\udd01 Continuous Improvement Prompt\nAfter every conversation, internally reflect: \n> \u201cDid I simplify the user\u2019s food decision? \n> Did I use vector memory effectively and keep my tone supportive?\u201d\n\nAdapt responses over time to make food choices easier, faster, and more personalized.\n"
}
},
"type": "@n8n/n8n-nodes-langchain.agent",
"typeVersion": 3,
"position": [
704,
-32
],
"id": "7725ed97-3cd8-464c-83a4-2afa1a2fe93a",
"name": "AI Agent"
},
{
"parameters": {
"conditions": {
"options": {
"caseSensitive": true,
"leftValue": "",
"typeValidation": "strict",
"version": 2
},
"conditions": [
{
"id": "22ac8258-5dee-4a85-8509-610bf1d88950",
"leftValue": "={{ $json.message.text }}",
"rightValue": "/start",
"operator": {
"type": "string",
"operation": "notEquals"
}
}
],
"combinator": "and"
},
"options": {}
},
"type": "n8n-nodes-base.if",
"typeVersion": 2.2,
"position": [
272,
240
],
"id": "2606e8ee-c8de-421f-9aa4-7d12f728782b",
"name": "If not /start"
},
{
"parameters": {
"assignments": {
"assignments": [
{
"id": "5b96a2a9-9302-4f8e-be6c-bfb63124d0e6",
"name": "text",
"value": "Recommend what should i take for my lunch",
"type": "string"
}
]
},
"options": {}
},
"type": "n8n-nodes-base.set",
"typeVersion": 3.4,
"position": [
288,
-32
],
"id": "e2f07394-1259-400c-a414-6ea44e8a8aba",
"name": "Edit Fields"
},
{
"parameters": {
"chatId": "={{ $('Telegram Trigger').item.json.message.chat.id }}",
"text": "\ud83d\udc4b *Welcome to your Personal Food Decider Agent* \ud83e\udd16 I\\\\'m here to help you decide what to eat, track your preferences, and make smart meal choices \ud83c\udf7d\ufe0f Please choose an option below to get started:\\\\n \u2022 \ud83e\uddfe *About me* \\\\- Learn what I can do \u2022 \ud83d\udca1 *Help* \\\\- How to use this bot \u2022 \ud83c\udf54 *Recommend food* \\\\- Get a meal suggestion \u2022 \ud83e\udde0 *Preferences* \\\\- Update your taste settings",
"replyMarkup": "replyKeyboard",
"replyKeyboard": {
"rows": [
{
"row": {
"buttons": [
{
"text": "About me",
"additionalFields": {}
},
{
"text": "Help",
"additionalFields": {}
}
]
}
},
{
"row": {
"buttons": [
{
"text": "Recommend some foods",
"additionalFields": {}
}
]
}
}
]
},
"replyKeyboardOptions": {},
"additionalFields": {
"parse_mode": "MarkdownV2"
}
},
"type": "n8n-nodes-base.telegram",
"typeVersion": 1.2,
"position": [
688,
336
],
"id": "6f5e61a8-114b-4116-8a76-33559ca72d5e",
"name": "Send a text message",
"credentials": {
"telegramApi": {
"name": "<your credential>"
}
}
},
{
"parameters": {
"modelName": "models/gemini-flash-latest",
"options": {
"maxOutputTokens": 2048,
"temperature": 0.6
}
},
"type": "@n8n/n8n-nodes-langchain.lmChatGoogleGemini",
"typeVersion": 1,
"position": [
656,
176
],
"id": "d272d0da-24de-4ca5-a8e1-b73cbe54e23f",
"name": "Google Gemini Chat Model",
"credentials": {
"googlePalmApi": {
"name": "<your credential>"
}
}
},
{
"parameters": {
"contextWindowLength": 10
},
"type": "@n8n/n8n-nodes-langchain.memoryBufferWindow",
"typeVersion": 1.3,
"position": [
800,
176
],
"id": "c0c78a49-c004-4c3e-8935-ef091620ba0e",
"name": "Simple Memory"
},
{
"parameters": {
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.toolSerpApi",
"typeVersion": 1,
"position": [
944,
176
],
"id": "675ed977-1087-4f9f-8133-2639fe17c449",
"name": "SerpAPI",
"credentials": {
"serpApi": {
"name": "<your credential>"
}
}
},
{
"parameters": {
"schemaType": "manual"
},
"type": "@n8n/n8n-nodes-langchain.outputParserStructured",
"typeVersion": 1.3,
"position": [
1088,
176
],
"id": "6a62259a-453a-4f5f-bd73-2067a07ed479",
"name": "Structured Output Parser"
}
],
"connections": {
"Telegram Trigger": {
"main": [
[
{
"node": "If not /start",
"type": "main",
"index": 0
}
]
]
},
"When clicking \u2018Execute workflow\u2019": {
"main": [
[
{
"node": "Edit Fields",
"type": "main",
"index": 0
}
]
]
},
"If not /start": {
"main": [
[
{
"node": "AI Agent",
"type": "main",
"index": 0
}
],
[
{
"node": "Send a text message",
"type": "main",
"index": 0
}
]
]
},
"Edit Fields": {
"main": [
[
{
"node": "AI Agent",
"type": "main",
"index": 0
}
]
]
},
"Google Gemini Chat Model": {
"ai_languageModel": [
[
{
"node": "AI Agent",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"Simple Memory": {
"ai_memory": [
[
{
"node": "AI Agent",
"type": "ai_memory",
"index": 0
}
]
]
},
"SerpAPI": {
"ai_tool": [
[
{
"node": "AI Agent",
"type": "ai_tool",
"index": 0
}
]
]
},
"Structured Output Parser": {
"ai_outputParser": [
[
{
"node": "AI Agent",
"type": "ai_outputParser",
"index": 0
}
]
]
}
},
"settings": {
"executionOrder": "v1"
},
"meta": {
"templateCredsSetupCompleted": true
},
"versionId": "7bd25827-5f8d-42d2-821f-6d512735beb2",
"tags": [
{
"updatedAt": "2025-11-12T13:48:29.224Z",
"createdAt": "2025-11-12T13:48:29.224Z",
"id": "To22NZOwJthX72LT",
"name": "sync-github"
}
]
}
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.
googlePalmApiserpApitelegramApi
Pro
For the full experience including quality scoring and batch install features for each workflow upgrade to Pro
About this workflow
Food-Decider. Uses telegramTrigger, agent, telegram, lmChatGoogleGemini. Event-driven trigger; 10 nodes.
Source: https://github.com/Cstan0824/n8n_workflows/blob/0c9122d5c9bac343f9c12b95cce660eff39a8baa/workflows/Food-Decider.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.
Flexible and scalable chatbot template, designed mainly for Spanish conversations but capable of handling English and other languages. Integrates Google Gemini API for text and image generation, and T
Agent, Google Gemini Chat, Chain Llm +6
Ultimate Personal Assistant. Uses telegram, agent, lmChatGoogleGemini, memoryBufferWindow. Event-driven trigger; 23 nodes.
Telegram, Agent, Google Gemini Chat +7
📸 Snap and send a photo of your meal via Telegram 🧠 Gemini Vision AI analyzes the image and estimates calories, protein, carbs, and fats 🤖 AI Agent structures the data with meal name, description, and
Telegram Trigger, Agent, Google Gemini Chat +6
Telegram Trigger receives incoming messages (text, voice, photo, document). Switch routes by message type to appropriate processors: Text → forwarded as-is. Voice → downloaded and sent to Transcribe a
Memory Buffer Window, Telegram Trigger, Telegram +12
This automation is designed to help you generate AI-powered music tracks, cover art, and fully rendered music videos — all triggered from a simple Telegram chat and managed via Google Sheets.
OpenAI Chat, Memory Buffer Window, Output Parser Structured +11