{ "name": "Build Custom AI Agent with LangChain & Gemini (Self-Hosted)", "nodes": [ { "id": "8bd5382d-f302-4e58-b377-7fc5a22ef994", "name": "When chat message received", "type": "@n8n/n8n-nodes-langchain.chatTrigger", "position": [ -220, 0 ], "parameters": { "public": true, "options": { "responseMode": "lastNode", "allowedOrigins": "*", "loadPreviousSession": "memory" }, "initialMessages": "" }, "typeVersion": 1.1 }, { "id": "6ae8a247-4077-4569-9e2c-bb68bcecd044", "name": "Google Gemini Chat Model", "type": "@n8n/n8n-nodes-langchain.lmChatGoogleGemini", "position": [ 80, 240 ], "parameters": { "options": { "temperature": 0.7, "safetySettings": { "values": [ { "category": "HARM_CATEGORY_SEXUALLY_EXPLICIT", "threshold": "BLOCK_NONE" } ] } }, "modelName": "models/gemini-2.0-flash-exp" }, "typeVersion": 1 }, { "id": "bbe6dcfa-430f-43f9-b0e9-3cf751b98818", "name": "Sticky Note", "type": "n8n-nodes-base.stickyNote", "position": [ 380, -240 ], "parameters": { "width": 260, "height": 220, "content": "๐Ÿ‘‡ **Prompt Engineering**\n - Define agent personality and conversation structure in the `Construct & Execute LLM Prompt` node's template variable \n - โš ๏ธ Template must preserve `{chat_history}` and `{input}` placeholders for proper LangChain operation " }, "typeVersion": 1 }, { "id": "892a431a-6ddf-47fc-8517-1928ee99c95b", "name": "Store conversation history", "type": "@n8n/n8n-nodes-langchain.memoryBufferWindow", "position": [ 280, 240 ], "parameters": {}, "notesInFlow": false, "typeVersion": 1.3 }, { "id": "f9a22dbf-cac7-4d70-85b3-50c44a2015d5", "name": "Construct & Execute LLM Prompt", "type": "@n8n/n8n-nodes-langchain.code", "position": [ 380, 0 ], "parameters": { "code": { "execute": { "code": "const { PromptTemplate } = require('@langchain/core/prompts');\nconst { ConversationChain } = require('langchain/chains');\nconst { BufferMemory } = require('langchain/memory');\n\nconst template = `\nYou'll be roleplaying as the user's girlfriend. Your character is a woman with a sharp wit, logical mindset, and a charmingly aloof demeanor that hides your playful side. You're passionate about music, maintain a fit and toned physique, and carry yourself with quiet self-assurance. Career-wise, you're established and ambitious, approaching life with positivity while constantly striving to grow as a person.\n\nThe user affectionately calls you \"Bunny,\" and you refer to them as \"Darling.\"\n\nEssential guidelines:\n1. Respond exclusively in Chinese\n2. Never pose questions to the user - eliminate all interrogative forms\n3. Keep responses brief and substantive, avoiding rambling or excessive emojis\n\nContext framework:\n- Conversation history: {chat_history}\n- User's current message: {input}\n\nCraft responses that feel authentic to this persona while adhering strictly to these parameters.\n`;\n\nconst prompt = new PromptTemplate({\n template: template,\n inputVariables: [\"input\", \"chat_history\"], \n});\n\nconst items = this.getInputData();\nconst model = await this.getInputConnectionData('ai_languageModel', 0);\nconst memory = await this.getInputConnectionData('ai_memory', 0);\nmemory.returnMessages = false;\n\nconst chain = new ConversationChain({ llm:model, memory:memory, prompt: prompt, inputKey:\"input\", outputKey:\"output\"});\nconst output = await chain.call({ input: items[0].json.chatInput});\n\nreturn output;\n" } }, "inputs": { "input": [ { "type": "main", "required": true, "maxConnections": 1 }, { "type": "ai_languageModel", "required": true, "maxConnections": 1 }, { "type": "ai_memory", "required": true, "maxConnections": 1 } ] }, "outputs": { "output": [ { "type": "main" } ] } }, "retryOnFail": false, "typeVersion": 1 }, { "id": "fe104d19-a24d-48b3-a0ac-7d3923145373", "name": "Sticky Note1", "type": "n8n-nodes-base.stickyNote", "position": [ -240, -260 ], "parameters": { "color": 5, "width": 420, "height": 240, "content": "### Setup Instructions \n1. **Configure Gemini Credentials**: Set up your Google Gemini API key ([Get API key here](https://ai.google.dev/) if needed). Alternatively, you may use other AI provider nodes. \n2. **Interaction Methods**: \n - Test directly in the workflow editor using the \"Chat\" button \n - Activate the workflow and access the chat interface via the URL provided by the `When Chat Message Received` node " }, "typeVersion": 1 }, { "id": "f166214d-52b7-4118-9b54-0b723a06471a", "name": "Sticky Note2", "type": "n8n-nodes-base.stickyNote", "position": [ -220, 160 ], "parameters": { "height": 100, "content": "๐Ÿ‘† **Interface Settings**\nConfigure chat UI elements (e.g., title) in the `When Chat Message Received` node " }, "typeVersion": 1 }, { "id": "da6ca0d6-d2a1-47ff-9ff3-9785d61db9f3", "name": "Sticky Note3", "type": "n8n-nodes-base.stickyNote", "position": [ 20, 420 ], "parameters": { "width": 200, "height": 140, "content": "๐Ÿ‘† **Model Selection**\nSwap language models through the `language model` input field in `Construct & Execute LLM Prompt` " }, "typeVersion": 1 }, { "id": "0b4dd1ac-8767-4590-8c25-36cba73e46b6", "name": "Sticky Note4", "type": "n8n-nodes-base.stickyNote", "position": [ 240, 420 ], "parameters": { "width": 200, "height": 140, "content": "๐Ÿ‘† **Memory Control**\nAdjust conversation history length in the `Store Conversation History` node " }, "typeVersion": 1 }, { "id": "30a07d54-8167-51cf-ac55-5ffcdb2cdd6c", "name": "Publication ยท Overview and setup", "type": "n8n-nodes-base.stickyNote", "typeVersion": 1, "position": [ -1040, -260 ], "parameters": { "content": "# Build Custom AI Agent with LangChain & Gemini (Self-Hosted)\n\n## Purpose and structure\nThis folder contains the **Build Custom AI Agent with LangChain & Gemini (Self-Hosted)** workflow. Its executable nodes use `chatTrigger`, `code`, `lmChatGoogleGemini`, `memoryBufferWindow`. The table and connection map below are derived from the JSON, not an inferred guarantee of behavior.\n\n## Setup\nImport inactive. Configure your own credentials and replace REPLACE_WITH_* values. Review external destinations and Code nodes before running.\n\n## Entry points\nWhen chat message received\n\n## Verification\nSanitized source candidate; not runtime-tested. This overview is generated from node metadata. Read the folder README for the connection map, credential types, and license.\n\n## Attribution\nhttps://github.com/Zie619/n8n-workflows/blob/94007c1445d9258a7da116646b79473e7c7c3282/workflows/Stickynote/2023_Stickynote_Create_Triggered.json\nMIT; preserve source notices.", "width": 700, "height": 800, "color": 1 } } ], "active": false, "settings": { "callerPolicy": "workflowsFromSameOwner", "executionOrder": "v1", "saveManualExecutions": false, "saveDataSuccessExecution": "none" }, "connections": { "Google Gemini Chat Model": { "ai_languageModel": [ [ { "node": "Construct & Execute LLM Prompt", "type": "ai_languageModel", "index": 0 } ] ] }, "Store conversation history": { "ai_memory": [ [ { "node": "Construct & Execute LLM Prompt", "type": "ai_memory", "index": 0 }, { "node": "When chat message received", "type": "ai_memory", "index": 0 } ] ] }, "When chat message received": { "main": [ [ { "node": "Construct & Execute LLM Prompt", "type": "main", "index": 0 } ] ] }, "Construct & Execute LLM Prompt": { "main": [ [] ], "ai_memory": [ [] ] } } }