mirror of
https://github.com/langbot-app/LangBot.git
synced 2026-11-01 13:25:18 +08:00
RockChinQ
9d2042478a
git-subtree-dir: docs git-subtree-mainline:0582d1b3d4git-subtree-split:51dddcb3f8
187 lines
7.1 KiB
Plaintext
187 lines
7.1 KiB
Plaintext
---
|
|
title: "n8n と LangBot でマルチプラットフォーム AI Chatbot を構築"
|
|
description: "How I Built a Multi-Platform AI Chatbot with n8n and LangBot"
|
|
---
|
|
|
|
この記事は LangBot Blog から同期されています。 [Blog の原文を読む](https://langbot.app/en/blog/n8n-multi-platform-ai-chatbot).
|
|
|
|
公開日：2025-12-03 · 著者：LangBot Team
|
|
|
|
> この記事の日本語版はまだ公開されていないため、英語版を掲載しています。
|
|
|
|
|
|
|
|
Connecting n8n's visual workflow automation with LangBot's multi-platform bot framework creates a powerful, code-free way to deploy AI chatbots across QQ, WeChat, Discord, Telegram, Slack, and more. This tutorial shows you how to integrate these tools in minutes.
|
|
|
|
## What You'll Need
|
|
|
|
- Python 3.8+ installed
|
|
- Node.js 18+ installed
|
|
- npm or npx available
|
|
- 15 minutes of your time
|
|
|
|
## Deploy LangBot in 3 Commands
|
|
|
|
LangBot is a production-ready bot framework that connects to multiple messaging platforms and AI services including Dify, FastGPT, Coze, OpenAI, Claude, and Gemini. Deploy it instantly:
|
|
|
|
```bash
|
|
cd your-workspace
|
|
mkdir -p langbot-instance && cd langbot-instance
|
|
uvx langbot@latest
|
|
```
|
|
|
|
On first launch, LangBot initializes automatically. Open `http://127.0.0.1:5300` in your browser.
|
|
|
|
|
|
|
|
Register your admin account when prompted. You'll land on the dashboard where you can manage bots, models, pipelines, and integrations.
|
|
|
|
|
|
|
|
## Set Up n8n Workflow Automation
|
|
|
|
n8n is an open-source automation platform with 400+ integrations and powerful AI capabilities. Launch it locally:
|
|
|
|
```bash
|
|
cd your-workspace
|
|
mkdir -p n8n-data
|
|
export N8N_USER_FOLDER=$(pwd)/n8n-data
|
|
npx n8n
|
|
```
|
|
|
|
Visit `http://127.0.0.1:5678` and create your owner account.
|
|
|
|
|
|
|
|
## Build Your AI Workflow
|
|
|
|
Create a new workflow in n8n. You'll need two essential nodes:
|
|
|
|
|
|
|
|
### Add the Webhook Trigger
|
|
|
|
Click "+" on the canvas and add a **Webhook** node. Configure it:
|
|
|
|
- **HTTP Method**: POST
|
|
- **Response Mode**: Streaming (enables real-time chat responses)
|
|
- **Authentication**: None (adjust for production)
|
|
|
|
|
|
|
|
### Add the AI Agent
|
|
|
|
Press Tab, navigate to the "AI" category, and select **AI Agent**.
|
|
|
|
**Configure the Chat Model**: Click "Chat Model" and choose "OpenAI Chat Model". Add your credentials:
|
|
|
|
- **API Key**: Your OpenAI API key (or compatible service key)
|
|
- **Base URL**: For OpenAI alternatives like Claude, Gemini, or local models, update to your provider's endpoint
|
|
|
|
**Critical Step - Fix the Prompt Source**: By default, the AI Agent expects a Chat Trigger node, which won't work with webhooks. Here's how to fix it:
|
|
|
|
1. Find "Source for Prompt (User Message)" in the AI Agent settings
|
|
2. Change from "Connected Chat Trigger Node" to "Define below"
|
|
3. Switch to "Expression" mode
|
|
4. Enter: `{{ $json.body }}`
|
|
|
|
This expression pulls the user's message from the webhook request body.
|
|
|
|
|
|
|
|
### Activate and Get Your Webhook URL
|
|
|
|
Save the workflow and toggle the activation switch (top-right). Switch to the "Production URL" tab and copy the webhook URL:
|
|
|
|
```
|
|
http://localhost:5678/webhook/{your-webhook-id}
|
|
```
|
|
|
|
## Connect LangBot to n8n
|
|
|
|
Back in the LangBot dashboard, navigate to **Pipelines** and click the default "ChatPipeline".
|
|
|
|
|
|
|
|
Switch to the **AI** tab and select "n8n Workflow API" from the Runner dropdown. Configure:
|
|
|
|
- **Webhook URL**: Paste your n8n production webhook URL
|
|
- **Authentication Type**: None (match your n8n webhook settings)
|
|
- **Timeout**: 120 seconds
|
|
- **Output Key**: response
|
|
|
|
Click **Save**.
|
|
|
|
## Test It Out
|
|
|
|
In the Pipeline editor, click "Debug Chat" on the left sidebar. Send a test message like "What is LangBot?"
|
|
|
|
If everything works, you'll see LangBot send the message to n8n, where the AI Agent processes it and streams back a response.
|
|
|
|
## Troubleshooting
|
|
|
|
**Error: "Expected to find the prompt in an input field called 'chatInput'"**
|
|
|
|
This means the AI Agent is still configured for a Chat Trigger node. Fix it:
|
|
|
|
1. Open the AI Agent configuration
|
|
2. Set "Source for Prompt (User Message)" to "Define below"
|
|
3. Switch to Expression mode
|
|
4. Enter: `{{ $json.body }}`
|
|
|
|
**Test Your Webhook Directly**
|
|
|
|
Verify the webhook works with curl:
|
|
|
|
```bash
|
|
curl -X POST http://localhost:5678/webhook/your-webhook-id \
|
|
-H "Content-Type: application/json" \
|
|
-d '{"body": "Hello, can you introduce yourself?"}'
|
|
```
|
|
|
|
You should receive streaming JSON with the AI's response.
|
|
|
|
## How the Integration Works
|
|
|
|
Here's the complete flow:
|
|
|
|
1. User sends a message via QQ, WeChat, Discord, Telegram, Slack, LINE, or any LangBot-supported platform
|
|
2. LangBot's Pipeline receives the message and calls the n8n Workflow API
|
|
3. n8n's Webhook node captures the request and passes it to the AI Agent
|
|
4. The AI Agent uses OpenAI, Claude, Gemini, or your configured LLM to generate a response
|
|
5. n8n streams the response back to LangBot
|
|
6. LangBot delivers the response to the user on their original platform
|
|
|
|
## Why This Combination Works
|
|
|
|
**LangBot + n8n** unlocks powerful capabilities:
|
|
|
|
1. **No-Code AI Logic**: Design conversation flows visually in n8n without touching code
|
|
2. **Multi-Platform Reach**: Deploy the same AI across QQ, WeChat, Discord, Telegram, Slack, LINE, DingTalk, and Lark simultaneously
|
|
3. **Flexible AI Models**: Swap between OpenAI GPT, Anthropic Claude, Google Gemini, Coze, Dify, local models, and more
|
|
4. **Rich Integrations**: Connect n8n's 400+ integrations - databases, APIs, Notion, Airtable, Google Sheets, Slack, and beyond
|
|
5. **Tool-Calling Agents**: AI Agent can trigger n8n tools like database queries, API calls, or custom functions
|
|
6. **Workflow Extensions**: Add preprocessing, content moderation, logging, or custom business logic
|
|
|
|
**Perfect For**:
|
|
- Enterprise customer support bots
|
|
- Knowledge base Q&A systems
|
|
- Multi-platform community management
|
|
- Task automation assistants
|
|
- Unified chat interfaces for teams
|
|
|
|
## Next Steps
|
|
|
|
Extend your bot further:
|
|
|
|
- Integrate Dify or FastGPT for advanced RAG (retrieval-augmented generation)
|
|
- Add vector database nodes (Pinecone, Qdrant, Weaviate) for knowledge retrieval
|
|
- Connect business APIs for real-time data
|
|
- Implement conversation memory and context tracking
|
|
- Add content filtering and moderation workflows
|
|
- Use Langflow or Coze for additional AI orchestration
|
|
|
|
This integration gives you the flexibility of code-based AI frameworks like LangChain with the simplicity of visual workflow builders - all while reaching users across every major messaging platform.
|
|
|
|
Ready to deploy your multi-platform AI assistant? Start with LangBot and n8n today.
|