Proxy MCP Server

proxy-mcp

proxy-mcp is an MCP server that runs an explicit HTTP/HTTPS MITM proxy (L7). It captures requests/responses, lets you modify traffic in-flight (headers/bodies/mock/forward/drop), supports upstream proxy chaining, and records TLS fingerprints for connections to the proxy (JA3/JA4) plus optional upstream server JA3S. It also ships "interceptors" to route Chrome, CLI tools, Docker containers, and Android devices/apps through the proxy.

81 tools + 8 resources + 4 resource templates. Built on mockttp.

Boundaries

• Only sees traffic configured to route through it (not a network tap or packet sniffer)
• Spoofs outgoing JA3 + HTTP/2 fingerprint + header order (via impit — native Rust TLS impersonation), not JA4 (JA4 is capture-only)
• Can add, overwrite, or delete HTTP headers; outgoing header order can be controlled via fingerprint spoofing
• Returns its own CA certificate — does not expose upstream server certificate chains

Pairs well with CDP/Playwright

Use CDP/Playwright for browser internals (DOM, JS execution, localStorage, cookie jar), and proxy-mcp for wire-level capture/manipulation + replay. They complement each other:

A typical combo: launch Chrome via interceptor_chrome_launch (routes through proxy automatically), drive pages with Playwright/CDP, and use proxy-mcp to capture the wire traffic, inject headers, or spoof JA3 — all in the same session. For behavioral realism, use humanizer_* tools instead of Playwright's instant page.click()/page.type() — they dispatch human-like CDP Input.* events with natural timing curves.

Attach Playwright to proxy-launched Chrome:

1. Call proxy_start
2. Call interceptor_chrome_launch
3. Read proxy://chrome/primary (or call interceptor_chrome_cdp_info) to get cdp.httpUrl (Playwright) and cdp.browserWebSocketDebuggerUrl (raw CDP clients)
4. In Playwright:

   import { chromium } from "playwright";
   const browser = await chromium.connectOverCDP("http://127.0.0.1:<cdp-port>");
   

Proxy-safe built-in CDP flow (single-instance safe):

1. Call proxy_start
2. Call interceptor_chrome_launch
3. Call interceptor_chrome_devtools_attach with that target_id
4. Call interceptor_chrome_devtools_navigate with devtools_session_id
5. Call proxy_search_traffic --query "<hostname>" to confirm capture

Human-like input flow (bypasses bot detection):

1. Call proxy_start
2. Optionally enable fingerprint spoofing: proxy_set_fingerprint_spoof --preset chrome_136
3. Call interceptor_chrome_launch --url "https://example.com" (stealth mode auto-enabled when spoofing)
4. Use humanizer_move / humanizer_click / humanizer_type / humanizer_scroll with the target_id
5. Use humanizer_idle between actions to maintain natural presence

HTTP Proxy Configuration

1) Start proxy and get endpoint

proxy_start

Use the returned port and endpoint http://127.0.0.1:.

2) Browser setup (recommended: interceptor)

Use the Chrome interceptor so proxy flags and cert trust are configured automatically:

interceptor_chrome_launch --url "https://example.com"

Then bind DevTools safely to that same target:

interceptor_chrome_devtools_attach --target_id "chrome_"
interceptor_chrome_devtools_navigate --devtools_session_id "devtools_<id>" --url "https://apify.com"

3) Browser setup (manual fallback)

If launching Chrome manually, pass proxy flag yourself:

google-chrome --proxy-server="http://127.0.0.1:"

4) CLI/process setup

Route any process through proxy-mcp by setting proxy env vars:

export HTTP_PR