slothproxy

slothproxy is a proxy server that adds configurable latency to requests, useful for testing how clients behave on slow connections. It supports two modes:

Reverse proxy mode

Forwards all incoming requests to a specific upstream server. Clients make normal requests to slothproxy's address.

# All requests to localhost:8080 are forwarded to api.example.com with 500ms latency
slothproxy serve --target http://api.example.com --latency 500ms

Forward proxy mode

Acts as a standard HTTP proxy. Clients configure their HTTP_PROXY to point at slothproxy.

# Start the forward proxy with 1s latency
slothproxy serve --latency 1s

# In another terminal
curl --proxy http://localhost:8080 http://example.com

Chaining with mitmproxy

Use --forward in forward proxy mode to chain through an upstream proxy like mitmproxy, giving you both latency simulation and traffic inspection:

slothproxy serve --latency 500ms --port 9999 --forward localhost:8080

Clients point at localhost:9999 -> slothproxy adds latency -> mitmproxy on localhost:8080 inspects traffic -> internet.

Selective throttling

Use --match to only throttle requests matching a regex pattern. Unmatched requests pass through at full speed.

# Only throttle API requests
slothproxy serve --target http://localhost:3000 --latency 2s --match "/api/"

# Only throttle a specific domain
slothproxy serve --latency 500ms --match "slow\.example\.com"

In forward proxy mode with HTTPS, --match is matched against the host:port (e.g. example.com:443) since the full URL isn't visible during CONNECT tunnel setup.

How latency works

The --latency value represents the minimum total request time, not additional delay. If you specify --latency 5s and the upstream responds in 2s, slothproxy sleeps for the remaining 3s. If the upstream takes longer than the specified latency, no extra delay is added.

All options

slothproxy serve [flags]

Flags:
  -f, --forward string     upstream proxy to forward traffic to in forward proxy mode
  -l, --latency duration   latency to add to each request (e.g. 200ms, 1s)
  -m, --match string       regex pattern to select which URLs get latency applied
  -p, --port int           port to listen on (default 8080)
  -t, --target string      upstream server URL for reverse proxy mode
      --verbose            enable verbose logging