Create Your First AI Agent

1. Workflow Overview

This workflow, titled "Build Your First Conversational AI Chatbot Agent," is designed to create an interactive conversational AI agent using Google Gemini (PaLM/Gemini Pro) via LangChain integration within n8n. It targets users seeking to build a customizable AI chatbot that can:

Receive chat messages from external chat platforms (via webhook)
Process these messages using Google Gemini AI models
Optionally fetch and simplify Google Docs content referenced in the chat
Log the conversation (user inputs and AI responses) to a Google Sheet for persistent storage and later analysis
Optionally send AI responses back to a Telegram chat for notification or direct user feedback

Logical Blocks:

1.1 Input Reception & Trigger: Listens and receives incoming chat messages.
1.2 AI Processing: Uses LangChain-connected Google Gemini AI to generate contextual replies.
1.3 Optional Google Docs Handling: Retrieves and simplifies Google Docs documents if referenced.
1.4 Logging & Output: Stores conversation data in Google Sheets and optionally sends replies to Telegram.
1.5 Auxiliary Nodes: Sticky notes for documentation, instructions, and optional configurations.

2. Block-by-Block Analysis

1.1 Input Reception & Trigger

Overview: This block captures incoming chat messages from external sources via a webhook and triggers the workflow.
Nodes Involved: Chat message
Node Details:
- Chat message
  - Type & Role: LangChain Chat Trigger node; acts as a webhook listener for incoming chat messages.
  - Configuration:
    - Public webhook enabled, allowing external chat systems to send messages.
    - Initial message configured as "Hi Nani! 👋" (sent to users when the chat starts).
  - Expressions/Variables: Captures user chat input as chatInput.
  - Inputs: External webhook (HTTP request).
  - Outputs: Passes user message JSON to the AI Agent node.
  - Edge cases: Webhook URL must be publicly accessible; malformed or missing chat input could cause failures.
  - Version: 1.1

1.2 AI Processing

Overview: Uses LangChain’s AI Agent node integrated with Google Gemini AI to generate intelligent, user-friendly replies to chat messages.
Nodes Involved: AI Agent, Gemini Chat, Gemini, Request
Node Details:
- AI Agent
  - Type & Role: LangChain Agent node responsible for orchestrating AI processing and tool usage.
  - Configuration: Uses default options; connects to AI language model and external tools.
  - Inputs: Receives message from Chat message node and tool outputs (Gemini, Docs, Request nodes).
  - Outputs: Sends AI-generated output to logging and Telegram nodes.
  - Edge cases: Requires valid API credentials; failure in any connected tool may affect output.
  - Version: 2.1
- Gemini Chat
  - Type & Role: LangChain Google Gemini Chat language model node.
  - Configuration: Uses Google Gemini API credentials; no additional options set.
  - Inputs: Connected as AI language model for AI Agent.
  - Outputs: Provides AI-generated chat responses.
  - Edge cases: API quota limits, network errors, or invalid modelId could cause failures.
  - Version: 1
- Gemini
  - Type & Role: LangChain Google Gemini tool node.
  - Configuration: Model ID set to models/gemini-2.5-flash; pre-message instructs the model to deliver user-friendly replies avoiding robotic language.
  - Inputs: Used as a tool by AI Agent.
  - Outputs: Provides processed text to AI Agent.
  - Edge cases: Same as Gemini Chat; API credential or usage limits.
  - Version: 1
- Request
  - Type & Role: HTTP Request node.
  - Configuration: Set to call URL https://google.cm/ (likely placeholder or test).
  - Inputs: Used as a tool by AI Agent.
  - Outputs: Provides HTTP response to AI Agent.
  - Edge cases: Network failure, invalid URL.
  - Version: 4.2

1.3 Optional Google Docs Handling

Overview: Retrieves and simplifies Google Docs content if the user's message references a Google Docs document.
Nodes Involved: Docs
Node Details:
- Docs
  - Type & Role: Google Docs node to fetch and simplify document content.
  - Configuration:
    - Operation: get
    - Document URL or ID dynamically set using AI override expression.
    - Simplify option also controlled via AI override.
  - Inputs: Used as a tool by AI Agent.
  - Outputs: Supplies simplified document text for AI processing.
  - Credentials: Requires OAuth2 Google Docs API credentials.
  - Edge cases: Invalid document URL/ID, permission errors, token expiration.
  - Version: 2

1.4 Logging & Output

Overview: Logs the conversation (user question and AI reply) into Google Sheets and optionally sends the AI response back to Telegram.
Nodes Involved: Store in sheet, Crypto, Store in Your Chat
Node Details:
- Store in sheet
  - Type & Role: Google Sheets node appending conversation data.
  - Configuration:
    - Operation: append
    - Document ID and sheet name (gid=0) fixed to a specific Google Sheet.
    - Columns mapped: "replay message" ← AI output, "output data (Crypto)" ← processed crypto output.
  - Inputs: Receives processed data from Crypto node.
  - Outputs: None (final storage).
  - Credentials: Requires OAuth2 Google Sheets credentials.
  - Edge cases: Sheet permissions, rate limits, malformed data.
  - Version: 4.6
- Crypto
  - Type & Role: Crypto node that processes AI output (could be for encryption or hashing).
  - Configuration: Input value set to AI generated output.
  - Inputs: Receives AI output from AI Agent.
  - Outputs: Sends processed data to Store in sheet.
  - Edge cases: Invalid input data may cause errors.
  - Version: 1
- Store in Your Chat
  - Type & Role: Telegram node sending messages to a Telegram chat.
  - Configuration:
    - Sends formatted message containing user question and AI reply.
    - Chat ID must be specified (placeholder "your chat ID").
  - Inputs: Receives AI output from AI Agent.
  - Outputs: None.
  - Credentials: Requires Telegram API credentials.
  - Edge cases: Invalid chat ID, Telegram API limits, network errors.
  - Version: 1.2

1.5 Auxiliary Nodes (Sticky Notes)

Overview: Provides documentation, usage instructions, reminders, and optional configurations visually within the workflow.
Nodes Involved: All sticky notes (Sticky Note, Sticky Note2, Sticky Note3, Sticky Note4, Sticky Note12, Sticky Note13)
Node Details:
- Sticky Note (large overview note)
  - Contains detailed overview, features, requirements, usage instructions, and helpful links.
- Sticky Note2 & Sticky Note3
  - Notes about tools and chat models customization.
- Sticky Note4
  - Describes optional Telegram notification and Google Sheets logging.
- Sticky Note12
  - Encourages activation and sharing of the public chat URL.
- Sticky Note13
  - Explains AI agent capabilities and system message tuning.

3. Summary Table

Node Name	Node Type	Functional Role	Input Node(s)	Output Node(s)	Sticky Note
Chat message	@n8n/n8n-nodes-langchain.chatTrigger	Receives incoming chat messages	External webhook	AI Agent	Activate this workflow and share the public chat URL to let others talk to your AI Agent!
AI Agent	@n8n/n8n-nodes-langchain.agent	Orchestrates AI processing	Chat message, Gemini, Docs, Request	Store in Your Chat, Crypto	Your AI agent can: Receive messages, select tools, respond with live answers. Adjust system message behavior.
Gemini Chat	@n8n/n8n-nodes-langchain.lmChatGoogleGemini	AI language model for chat	AI Agent (ai_languageModel)	AI Agent	Chat modal can be changed; requires API key.
Gemini	@n8n/n8n-nodes-langchain.googleGeminiTool	AI tool for friendly replies	AI Agent (ai_tool)	AI Agent	Chat modal can be changed; requires API key.
Request	n8n-nodes-base.httpRequestTool	HTTP call tool	AI Agent (ai_tool)	AI Agent	Chat modal can be changed; requires API key.
Docs	n8n-nodes-base.googleDocsTool	Fetches and simplifies Google Docs	AI Agent (ai_tool)	AI Agent	Tool can be added/changed; requires OAuth2 Google Docs API.
Store in sheet	n8n-nodes-base.googleSheets	Logs conversation in Google Sheets	Crypto	None	Optional: User can get notifications on Telegram and store in Google Sheets.
Crypto	n8n-nodes-base.crypto	Processes AI output (crypto)	AI Agent	Store in sheet
Store in Your Chat	n8n-nodes-base.telegram	Sends AI replies to Telegram chat	AI Agent	None	Optional: User can get notifications on Telegram and store in Google Sheets.
Sticky Note	n8n-nodes-base.stickyNote	Documentation & instructions	None	None	Contains workflow overview, features, requirements, usage instructions, and helpful community links.
Sticky Note2	n8n-nodes-base.stickyNote	Notes on tool customization	None	None	Tools can be changed but require API keys and data for replies.
Sticky Note3	n8n-nodes-base.stickyNote	Notes on chat model customization	None	None	Chat modal can be changed but requires API keys.
Sticky Note4	n8n-nodes-base.stickyNote	Optional Telegram notification info	None	None	Optional user notification on Telegram and Google Sheets logging.
Sticky Note12	n8n-nodes-base.stickyNote	Activation and sharing reminder	None	None	Activate workflow and share public chat URL.
Sticky Note13	n8n-nodes-base.stickyNote	AI agent capability explanation	None	None	Describes AI agent's receiving, tool selecting, and responding capabilities along with system message tuning.

4. Reproducing the Workflow from Scratch

Create Chat Trigger Node
- Add LangChain Chat Trigger node.
- Configure webhook as public.
- Set initial message: "Hi Nani! 👋".
- This node receives chat inputs from users.
Add LangChain AI Agent Node
- Add LangChain Agent node.
- Connect output of Chat Trigger node to input of AI Agent node.
- Leave options default; this node coordinates AI processing.
Add Gemini Chat Model Node
- Add LangChain Google Gemini Chat node.
- Configure credentials with Google Gemini (PaLM) API OAuth2 credentials.
- Connect this node as the AI language model input to the AI Agent node.
Add Gemini AI Tool Node
- Add LangChain Google Gemini Tool node.
- Set modelId to "models/gemini-2.5-flash".
- Add pre-message instruction: "Give me user user-friendly reply. Don't give me a robotic type relay."
- Configure credentials with Google Gemini (PaLM) API OAuth2 credentials.
- Connect this node as an AI tool input to the AI Agent node.
Add HTTP Request Tool Node
- Add HTTP Request node.
- Set URL to "https://google.cm/" (can be customized).
- Connect this node as an AI tool input to the AI Agent node.
Add Google Docs Node
- Add Google Docs node.
- Operation: get.
- Set Document URL dynamically to an expression that can be overridden via AI input.
- Set Simplify option dynamically (boolean) via AI override.
- Connect this node as an AI tool input to the AI Agent node.
- Configure credentials with Google Docs OAuth2 credentials.
Add Telegram Node for Chat Notification
- Add Telegram node.
- Configure credentials with Telegram API OAuth2.
- Set message text to include user question and AI reply using expressions.
- Specify the chatId for the Telegram chat to send messages.
- Connect output of AI Agent node to this node.
Add Crypto Node
- Add Crypto node.
- Set input value dynamically to AI output ({{ $json.output }}).
- Connect output of AI Agent node to this node.
Add Google Sheets Node
- Add Google Sheets node.
- Operation: append.
- Set Spreadsheet ID to desired Google Sheet document ID.
- Set Sheet name (e.g., gid=0).
- Define columns mapping:
  - "replay message" ← AI output
  - "output data (Crypto)" ← Crypto node output
- Connect output of Crypto node to this node.
- Configure credentials with Google Sheets OAuth2 credentials.
Connect AI Agent Outputs
- Connect main output of AI Agent node to Telegram node and Crypto node.
- Connect output of Crypto node to Google Sheets node.
Add Sticky Notes (Optional)
- For documentation and instructions, add sticky notes with relevant content to guide users.
Activate Workflow
- Ensure all credentials are valid and workflow is activated.
- Share the public webhook URL from Chat message node for integration with chat platforms.

5. General Notes & Resources

Note Content	Context or Link
Workflow listens for chat messages, uses Google Gemini AI (PaLM) via LangChain, fetches Google Docs, logs to Sheets	Overview and key features as documented in the main sticky note.
Requires OAuth2 credentials for Google APIs (Gemini, Docs, Sheets) and Telegram API credentials	Credential setup essential for functionality.
Telegram chat ID must be set to receive AI notification messages	Telegram node configuration requirement.
Public webhook URL from Chat Trigger node allows integration with multiple chat platforms (Telegram, Slack, etc.)	Enables external chat system connection.
Helpful community & author links:	devcodejourney.com, LinkedIn Shakil, WhatsApp Channel, Direct Chat
AI Agent can be extended with custom tools and models by modifying LangChain Agent's tool and model inputs	Flexible modular design for enhancements.
Google Docs document URL and simplify setting are dynamically handled via AI overrides	Allows dynamic document retrieval per user input.
Crypto node can be customized for encryption or hashing of outputs before storing	Security or integrity enhancement optional.
Initial greeting message can be customized in Chat Trigger node	Personalizes user experience.

This detailed reference fully describes the workflow structure, node configurations, connections, and usage instructions, enabling advanced users or automation agents to understand, reproduce, or modify it effectively.

18 KiB Raw Blame History Unescape Escape