Model Context Protocol
Model Context Protocol - The open standard for connecting AI models to tools and data
The Model Context Protocol (MCP) is an open standard created by Anthropic that standardizes how applications provide context to large language models (LLMs). SystemPrompt leverages MCP to create powerful voice-controlled AI interactions with persistent memory and tool capabilities.
What is MCP?
MCP acts like a "USB-C port" for AI applications - a standardized way to connect AI models to different data sources and tools:
- Defines how applications provide context to LLMs
- Enables tools and services to seamlessly integrate with language models
- Creates a structured information layer that persists between interactions
- Facilitates secure data access while respecting privacy boundaries
Core Capabilities
MCP provides several foundational capabilities that power SystemPrompt and other AI applications:
- Resources: Expose file-like data (API responses, documents, images) to LLMs
- Tools: Enable LLMs to execute functions and access services (with user approval)
- Prompts: Define reusable templates for common tasks and workflows
- Roots: Set boundaries for where servers can operate (filesystem paths, API endpoints)
- Sampling: Let servers request completions from LLMs, enabling agentic behaviors
Why SystemPrompt Uses MCP
SystemPrompt leverages MCP to provide several key advantages:
- Voice-First Design: Natural language voice commands with contextual awareness
- Tool Integration: Connect to Reddit, Gmail, and other services without coding
- Persistence: Maintain context across multiple sessions and devices
- Security: User-controlled permissions for all tool actions
- Extensibility: Add new capabilities through the growing MCP ecosystem
How MCP Works
MCP follows a client-server architecture where:
- Hosts (like SystemPrompt) initiate connections to MCP servers
- Clients maintain 1:1 connections with servers inside the host application
- Servers provide context, tools, and prompts to the clients
This architecture allows for modular, secure, and flexible AI integrations.
Learn more about the technical architecture and available integrations in the following sections.