MCP Servers
MCP (Model Context Protocol) servers extend your chatbot's capabilities by connecting it to external tools and services. When a visitor asks a question that requires an external tool, the chatbot automatically calls the appropriate MCP server.
MCP Servers are available on paid plans only.
Adding an MCP Server
Navigate to your chatbot's Tools tab, scroll to the MCP Servers section, and click Add MCP Server.
Configuration Fields
| Field | Description | Required |
|---|---|---|
| Name | A descriptive name for this MCP server (e.g. "Weather Service", "CRM Lookup"). | Yes |
| URL | The MCP server endpoint URL (e.g. https://mcp.example.org/sse). Must be a valid HTTP/HTTPS URL. | Yes |
| Description | Explain what this server does. This helps you and your team understand the server's purpose. | No |
| Headers (JSON) | Optional HTTP headers sent with every request, in JSON format (e.g. {"Authorization": "Bearer token123"}). Must be valid JSON. | No |
| Enabled | Whether this MCP server is active. Disable to temporarily stop using it without deleting the configuration. | Yes (default: On) |
Tool Toggles
After connecting an MCP server, WebChatAgent discovers the available tools it provides. Each tool can be individually enabled or disabled using toggle switches. This lets you control exactly which capabilities your chatbot has access to.
For example, if an MCP server provides both "search" and "delete" tools, you can enable "search" while keeping "delete" disabled.
How It Works
- A visitor asks a question that requires external data or actions
- The chatbot identifies the relevant MCP server and tool
- It sends a request to the MCP server with the necessary parameters
- The MCP server executes the action and returns results
- The chatbot uses the results to formulate its response
Example Use Cases
- CRM integration — Look up customer information by email or phone
- Inventory check — Query real-time stock levels for products
- Weather data — Fetch current weather conditions for a location
- Ticket system — Create or look up support tickets
- Database queries — Search internal databases for specific information
Best Practices
- Use descriptive names — Makes it easy to identify servers in the configuration
- Add descriptions — Document what each server does for your team
- Disable unused tools — Only enable the tools your chatbot actually needs
- Secure your endpoints — Use authentication headers to protect your MCP servers
- Test thoroughly — Verify that the MCP server responds correctly before enabling it in production