Foxhound is a compliance-grade observability platform for AI agent fleets, designed around tracing, replaying, and auditing every agent decision from tool call to business outcome. As an MCP server, it exposes application observability, tracing, prompt evaluation, dataset management, experiment management, budget enforcement, SLA monitoring, audit logging, and multi-tenant data isolation capabilities to AI assistants and MCP clients such as Claude Code, Cursor, and Windsurf. Foxhound integrates with TypeScript and Python agent environments, offers SDKs and a CLI, and supports self-hosted deployments for organizations with security and compliance requirements. It is aimed at enterprise AI teams, developers, and organizations needing secure, transparent oversight of AI agent behavior in production and development workflows.
Visit Foxhound's official website for product details and getting started.