Quick Start
Three steps to get GoLiveChat running. MySQL is the only external dependency.
Step 1: Configure MySQL
Edit config/mysql.json and set your MySQL host, port, user, password, and database name.
Step 2: Install & Import Database
./golivechat install
This command creates all required tables and populates initial data automatically.
Step 3: Start the Server
./golivechat server
Your live chat platform is now running. Access the admin panel at http://localhost:8080/main.
System Requirements
| CPU | 1 core minimum, 2+ recommended |
| RAM | 2 GB minimum |
| Database | MySQL 8.0+ |
| OS | Linux, macOS, Windows |
Embed the Chat Widget
After deployment, add this to your website:
<script type='text/javascript'>
(function(a,b,c,d){
let h=b.getElementsByTagName('head')[0];
let s=b.createElement('script');
s.type='text/javascript';
s.src=c+"/static/js/kefu-front.js";
s.onload=function(){d(c);};
h.appendChild(s);
})(window,document,"https://your-domain.com",function(u){
KEFU.init({
KEFU_URL:u,
KEFU_KEFU_ID:"your-agent-id",
KEFU_ENT:"your-tenant-id"
});
});
</script>
Configure AI
GoLiveChat supports multiple AI backends. Set your API key in the admin settings panel:
| OpenAI | GPT-3.5, GPT-4, GPT-4o — standard & Azure endpoints |
| Coze | Multi-model agent platform (Doubao, Kimi, Qwen, GLM-4) |
| Dify | Open-source AI platform — self-hosted or cloud |
| FastGPT | Knowledge-base Q&A with your documents |
Configure RAG Knowledge Base
To enable document-based AI answers:
- Start Qdrant vector database (
docker-compose up qdrant)
- Go to Admin → Knowledge Base → Upload Documents
- Upload PDF, Word, Excel, or TXT files
- Documents are automatically chunked, embedded, and indexed
- When visitors ask questions, the system retrieves relevant context before the AI responds