Mcp Agent Trace Inspector

Mcp Agent Trace Inspector provides local-first, privacy-preserving observability for AI agent workflows within the Model Context Protocol (MCP) ecosystem. It records every tool call, prompt transformation, latency, and token count in a local SQLite database, enabling secure, in-depth analysis of agent traces with no cloud dependence. Its features include step-by-step trace logging, persistent storage, token cost estimation, trace comparison, and an interactive HTML dashboard export, making it ideal for MCP-native environments where data privacy is paramount.