Architecture

MCP Architecture

The Model Context Protocol (MCP) follows a client-server architecture that standardizes communication between AI applications and external tools or data sources. This document provides a technical overview of how MCP works in the context of SystemPrompt.

Protocol Layer

The protocol layer handles message framing, request/response linking, and communication patterns:

• JSON-RPC 2.0: All MCP communication uses the JSON-RPC protocol
• Capability Negotiation: Clients and servers negotiate supported features
• Version Compatibility: Ensures backward compatibility between implementations
• Structured Messages: Standardized formats for all communications

Transport Layer

MCP supports multiple transport mechanisms for different use cases:

1. Stdio Transport

   • Uses standard input/output for communication
   • Ideal for local processes and system integration
   • Commonly used for desktop applications like SystemPrompt

2. HTTP with SSE Transport

   • Server-Sent Events for server-to-client communication
   • HTTP POST for client-to-server messages
   • Better for distributed systems and web applications

Communication Flow

The typical message flow in an MCP-enabled application like SystemPrompt:

1. Initialization

   • Client sends capabilities and version information
   • Server responds with supported features
   • Connection established with negotiated capabilities

2. Discovery

   • Client requests available tools, resources, or prompts
   • Server provides metadata about available capabilities
   • SystemPrompt UI displays available functions to user

3. Interaction

   • User requests an action (e.g., "Show my Reddit feed")
   • SystemPrompt determines which MCP server to use
   • Client sends appropriate request to server
   • Server executes operation and returns results
   • SystemPrompt displays results to user

SystemPrompt Implementation

Voice Processing Integration

SystemPrompt extends the MCP architecture with voice-specific components:

• WebRTC Voice Pipeline: Processes voice commands in real-time
• Intent Recognition: Maps voice commands to appropriate MCP tools
• Context Manager: Maintains conversation state across interactions
• Permission Management: Ensures user approval for sensitive operations

Cross-Device Synchronization

SystemPrompt implements the MCP specification with additional synchronization features:

• State Persistence: Conversation context synced across devices
• Tool Configuration: User tool preferences maintained universally
• Credential Management: Secure OAuth token handling for service integrations
• Preference Synchronization: User settings available on all devices

Security Model

The MCP architecture emphasizes security through several mechanisms:

• Human-in-the-Loop: User approval required for tool execution
• Sandboxed Environments: Isolated execution for third-party servers
• Permission Boundaries: Clearly defined access limitations
• Transport Security: Encryption for all communications
• Authentication: OAuth 2.0 for service authentication
• Audit Logging: Comprehensive activity tracking

Available Integrations

SystemPrompt leverages MCP to provide numerous integrations:

For implementation details on specific integrations, see our Integration documentation.