# AxLoop AI

> AxLoop AI is building local-first MCP fleet observability for AI agents across laptops, phones, IDEs, browser agents, embedded applications, edge devices, private infrastructure, and cloud runtimes.

## Core pages
- https://www.axloop.ai/mcp-fleet-observability — Technical definition and evaluation guide for MCP fleet observability.
- https://www.axloop.ai/product — Collector, fleet dashboard, and policy roadmap.
- https://www.axloop.ai/architecture — Local-first, federated, OpenTelemetry-native architecture.
- https://www.axloop.ai/company — Principles, intended users, and early access.

## Use cases
- https://www.axloop.ai/use-cases/shadow-mcp-discovery — Discover unapproved MCP servers and tools.
- https://www.axloop.ai/use-cases/mcp-tool-call-observability — Trace MCP calls across clients, servers, and dependencies.
- https://www.axloop.ai/use-cases/agent-cost-attribution — Attribute model and tool cost to workflows and teams.
- https://www.axloop.ai/use-cases/mcp-data-flow-monitoring — Map agent and MCP data-access paths with edge redaction.

## Comparisons
- https://www.axloop.ai/compare/server-apm-vs-mcp-observability — Server APM compared with client-edge MCP fleet observability.

## Product-status note
The collector is described as in development. The fleet dashboard is next. Policy and governance capabilities are planned. Illustrative sample data is labeled as such.