Whistle MCP Server – Overview

Whistle MCP Server bridges the gap between AI assistants and local network debugging by exposing a rich set of Whistle proxy operations through the Model Context Protocol. Rather than opening a browser-based interface or manually editing configuration files, developers can issue natural‑language commands to an AI assistant and have it create rules, manage groups, adjust proxy settings, or replay traffic—all in real time. This eliminates the friction of context switching and lets teams iterate faster on API design, security testing, or performance tuning.

The server translates MCP tool calls into Whistle commands. For example, a simple request can add a new rewrite or block rule; and toggle activation without touching the UI. Group operations let you bundle rules logically, mirroring how developers organize environments or feature flags. Value management adds dynamic variables that can be referenced across rules, enabling parameterized configurations. Proxy control covers the full spectrum of Whistle features—enabling/disabling HTTP/HTTPS interception, toggling HTTP/2 support, or switching the entire proxy on or off—so an AI can reconfigure a debugging session with a single prompt.

Key capabilities include:

• Rule lifecycle management: create, update, rename, delete, enable/disable, and bulk disable rules.
• Group orchestration: create, rename, delete groups; move rules in and out of groups.
• Value handling: add, update, rename, delete values, with group support for logical grouping of variables.
• Proxy toggles: start/stop the proxy, switch interception modes, and control protocol support.
• Request insight: list intercepted requests with URL filtering; replay captured traffic with custom parameters.

Real‑world scenarios benefit from this tight integration. QA teams can ask an AI to “enable the error‑mock rule for all POST endpoints” while a developer simultaneously tests downstream services. Security auditors might request “replay the last 50 requests to with altered headers” to validate input sanitization. API designers can prototype new endpoints by “create a rule that redirects to the staging server” and immediately see the effect in their browser, all without leaving the AI chat.

By exposing these operations over MCP, Whistle Server fits seamlessly into any AI‑driven workflow—whether on Claude Desktop, Raycast, or Cursor. Its declarative toolset empowers assistants to act as first‑class network engineers, reducing manual effort and accelerating the feedback loop in modern web development pipelines.