Csharp Sdk

The Model Context Protocol C# SDK is a library for building clients and servers that implement the Model Context Protocol to integrate AI tools and resources. It provides an AI tool integration framework and a multi-modal content handler to exchange text, images, and binary resources between AI models and external context providers.

The SDK utilizes a JSON-RPC communication library to manage bidirectional data exchange. It features a transport-agnostic communication layer that supports standard input and output, HTTP, and in-memory pipes, with specific integration for ASP.NET Core hosting.

The framework covers tool and prompt management, including schema-driven tool generation and attribute-based capability discovery. It supports resource management via URI-template mapping, asynchronous task orchestration, and secure authorization workflows through token exchange and identity assertion.

Features

AI Integration Protocols - Provides a standardized communication protocol for exchanging context, tools, and data between AI systems and software.

Model Context Protocol Servers - Implements the server-side framework for exposing tools, prompts, and data resources to AI hosts.

Tool Orchestration - Coordinates tool definitions and execution flows between AI models and clients using machine-readable schemas.

AI Tool Integrations - Converts discovered server tools into compatible formats for direct consumption by AI chat clients.

MCP Client Implementations - Provides a full client-side framework for connecting to MCP servers via stdio, SSE, or HTTP transport.

MCP Servers - Provides a framework for implementing MCP servers that expose tools, prompts, and data sources to AI models.

Resource Exposure - Exposes files, database records, and API responses as read-only resources for AI model context.

Model Context Protocol Clients - Provides the client-side integration layer allowing AI models to discover and execute remote tools.

Model Context Protocol Implementations - Provides a C# SDK for building clients and servers that implement the Model Context Protocol.

Context Sharing - Facilitates the passing of rich media like images and audio between AI clients and servers.

Multi-Modal Tool Outputs - Provides the ability to return rich media and images from tools for AI model analysis.

Dynamic Resource Templates - Uses URI templates to match and serve dynamic resources based on provided parameters.

MCP Resource Handlers - Provides handlers to retrieve data from servers as either plain text or raw byte blobs.

Prompt Embeddings - Supports inserting images, audio, and binary resources directly into AI prompt data blocks.

AI Client-Server Communication Protocols - Establishes bidirectional communication between AI applications and data servers using JSON-RPC over HTTP or stdio.

JSON-RPC Implementations - Implements a request-response and notification pattern using the standardized JSON-RPC format.

Protocol-Agnostic Transport Layers - Decouples protocol logic from the underlying data stream via abstractions for HTTP, stdio, and pipes.

AI Tool Schema Generation - Automatically converts C# method signatures into JSON schemas to make server functions callable by AI models.

Multi-Modal Handlers - Implements a handler for exchanging text, images, and binary resources between AI models and context providers.

AI Tool Discovery Schemas - Automatically converts C# method signatures into JSON schemas to make server functions discoverable by AI models.

Dynamic Tool Management - Allows updating the list of available tools at runtime and notifying connected clients to refresh their definitions.

Assembly-Based Discovery - Uses assembly scanning and custom markers to automatically register server tools and resources at runtime.

Transport Layer Connectivity - Connects clients to servers running as local child processes using standard input and output streams.

Tool Invocation Handlers - Enables querying remote servers for tool definitions and executing them through a client interface.

Prompt Templates - Supports the definition of parameterized, reusable message structures that AI clients can discover and compose.

Capability Registrations - Implements assembly scanning and attribute markers to automatically register server tools and resources.

Protocol Server Hosting - Enables the operation of protocol servers over HTTP to facilitate communication between models and external context.

Connection and Session Management - Coordinates data and command exchanges between clients and servers to maintain stable connection states.

MCP Transport Protocol Supports - Supports bidirectional communication between remote clients and servers via streamable HTTP or SSE.

Session State Management - Tracks client-server connectivity and capability negotiations via unique session identifiers.

Hosting Integrations - Integrates with ASP.NET Core to expose protocol capabilities over HTTP with session tracking.

modelcontextprotocolcsharp-sdk

Features

Star history