MCP Servers
Connect your chatbot to external tools and services using the Model Context Protocol standard.
What is MCP?
MCP (Model Context Protocol) is an open standard that lets chatbots communicate with external tools and services. With MCP, your bot can go beyond answering questions — it can perform actions, query real-time data, and integrate with your business systems.
What Can MCP Do?
- Query databases in real time
- Connect to external APIs
- Execute actions in other systems
- Retrieve live information from services
- Create automated workflows
Use Case Examples
Inventory lookup
The bot queries your inventory system to check if a product is in stock before answering the customer.
Real-time reservations
Connect to your booking system to show live availability and create reservations directly from chat.
CRM integration
Automatically register leads in Salesforce, HubSpot, or other CRMs when the bot captures contact information.
Order tracking
Let customers check their order status by connecting to your order management system.
Setup
- Go to AxiaChat AI → Settings → Advanced (or the MCP tab if available).
- Add the URL of your MCP server.
- Configure credentials if required.
- Save and test the connection.
Assigning MCP Tools to Bots
Each bot can have different MCP tools enabled:
- Go to AxiaChat AI → Bots and edit a bot.
- In the tools/capabilities section, enable the MCP tools you need.
- The bot will only use the tools assigned to it.
How It Works Automatically
- The user asks a question that requires external data
- The AI automatically detects it needs a tool
- It calls the appropriate MCP server
- Receives the response and integrates it into its answer
Everything is transparent to the end user — they see a natural response.
Security
- Isolated execution — Tools run in separate contexts
- No secret exposure — API keys are never sent to AI models
- Admin authorization — All tools require explicit enablement
- Execution logs — Every call is logged for auditing
Tool Logs
Review MCP tool usage at AxiaChat AI → Logs → Tools:
- Execution duration
- Output excerpts
- Error details (if any)
- Search and filter interface
Building Your Own MCP Server
If you have technical skills, you can build custom MCP servers for your specific needs. Refer to the Model Context Protocol documentation for server development details.