Features
Feature support for Model Context Protocol capabilities in Tambo
Tambo supports various Model Context Protocol features that enable rich interactions between your application and MCP servers. Below is a summary of currently supported features.
Supported Features
| Feature | Support Status | Description | Documentation |
|---|---|---|---|
| Tools | ✅ Supported | Call functions exposed by MCP servers to perform actions and retrieve data | Coming soon |
| Prompts | ✅ Supported | Use predefined prompt templates from MCP servers | View docs |
| Resources | ✅ Supported | Access and reference data sources exposed by MCP servers | View docs |
| Elicitations | ✅ Supported | Request additional input from users during tool execution with dynamic forms | View docs |
| Sampling | ✅ Supported | Allow MCP servers to request LLM completions through your application | View docs |
Feature Details
Tools
MCP tools are automatically discovered and made available to Tambo when you connect to an MCP server (either server-side or client-side). The AI can call these tools to perform actions like creating issues, fetching data, or interacting with external services.
Tools support rich content responses, including text, images, and other media types that are seamlessly passed to the AI.
Prompts
Prompts allow MCP servers to expose reusable prompt templates that can be inserted into your message input. Users can select from available prompts via a button or by typing "/" at the start of an empty text field.
Note: Prompt parameters are not yet supported. All prompts are inserted as-is without customization.
Resources
Resources allow MCP servers to expose data sources that users can reference in their conversations. When a user selects a resource, it's inserted with the syntax @<resourceUri>. Resources can be fetched on the server (for server-side MCP connections) or client (for client-side MCP connections) and are automatically formatted for the AI provider being used.
Elicitations
Elicitations allow MCP servers to pause during tool execution and request additional information from users. Tambo provides built-in UI components that automatically render dynamic forms based on the requested schema, handle validation, and return responses to the MCP server.
Sampling
Sampling enables MCP servers to request LLM completions through your application. This allows MCP tools to leverage AI capabilities for tasks like generating text, analyzing content, or making decisions based on context.
Note: Sampling currently only works with server-side MCP connections.
Connection Types
MCP features work with both connection types:
- Server-side MCP - Configured through the Tambo dashboard
- Client-side MCP - Connected directly from the browser
Different features may have different capabilities or requirements depending on the connection type. Check the specific documentation for each feature for details.