AI Chatbot Integration Documentation
Real-world implementation guides written from actual project experience. We've spent years building chatbots for different businesses, and these docs reflect what actually works—not just theoretical approaches.
Quick Start Guide
Get your first chatbot running in under an hour. We focus on the essential setup steps that matter most.
API Reference
Complete endpoint documentation with real response examples from production environments.
Best Practices
Patterns we've developed through building hundreds of conversation flows for actual clients.
Platform Integration Methods
Web Widget Integration
Add a single script tag to your site. That's it. The widget handles authentication, state management, and UI updates automatically. Works with React, Vue, or plain HTML.
REST API Access
Connect from any backend—Node, Python, PHP, whatever you're using. Standard JSON responses, clear error codes, and webhook support for real-time updates.
Custom Channel Builds
Want to deploy on Slack, Teams, or WhatsApp? We provide the base bot logic while you handle the specific platform connection. Documentation includes working examples for all major platforms.
Training Data Formats
Upload FAQs as CSV files or connect to your knowledge base directly. The system learns from your existing documentation—no need to restructure everything.
Technical Specifications
Everything your development team needs to know about our chatbot infrastructure and capabilities.
Response Performance
- Average API response: 180-240ms
- Widget load time: Under 2 seconds
- Concurrent user handling: 500+ per instance
- Message processing: 50 per second sustained
Data & Security
- Conversation data stored in Singapore
- AES-256 encryption at rest
- TLS 1.3 for all API traffic
- PDPA-compliant data handling
Language Support
- English with regional variations
- Chinese (Simplified and Traditional)
- Malay language processing
- Tamil basic query support
Integration Options
- JavaScript SDK for web deployment
- RESTful API with OpenAPI spec
- Webhook callbacks for events
- CSV/JSON bulk import tools
Mei Lin Tan
Integration Specialist
I've helped over 80 companies deploy our chatbot solutions. Most implementation questions I get have been asked before—so I put the answers here.
Implementation Walkthrough
Let me show you how a typical integration actually works. This is based on a retail client we set up in January 2025.
Initial Authentication Setup
You'll need an API key from your dashboard. Create one under Settings → API Access. Each key can be scoped to specific permissions—most clients start with read-write access for testing, then lock it down later.
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"message": "Hello", "session_id": "user_123"}'
The response includes the bot's reply and a conversation context token. Store that token—you'll need it for follow-up messages in the same conversation thread.
Widget Customization
The web widget comes with sensible defaults, but you'll probably want to adjust colors and position. Here's what most clients change first:
MentrixaChatbot.init({
apiKey: 'YOUR_API_KEY',
primaryColor: '#AA1BA3',
position: 'bottom-right',
greeting: 'Hi! How can I help you today?'
});
</script>
Common Pitfall
Don't initialize the widget before the DOM loads. Wrap your init call in a DOMContentLoaded listener if you're adding the script to your head section. Sounds obvious, but I've debugged this issue at least a dozen times.
Training Your Bot
The system needs examples to learn from. Upload a CSV with question-answer pairs, or point it at your help center URL. Processing usually takes 15-30 minutes depending on content volume.
Start with your top 20 customer questions. That covers most interactions for most businesses. You can always add more later—the bot learns incrementally.
Testing Before Launch
Use test mode to try conversations without logging them to your analytics. This way you can refine responses without messy data later. Switch to production mode once you're happy with how it handles queries.
Real Success Metric
Track your resolution rate—the percentage of conversations where users don't click "Talk to a human." Anything above 60% is solid for most industries. Don't obsess over hitting 90%+ unless you have extremely straightforward queries.
Need Implementation Help?
Our team can guide you through setup, answer technical questions, or handle the integration entirely. Most implementations take 2-3 days from start to finish.
Get Integration Support