rj

rj is a small CLI for printing the metadata of an HTTP response as JSON: status code, headers, protocol, and timing breakdown. It does not print the response body — for that, keep using curl. The two are complementary: curl for the body, rj for everything around it.

Install

brew install yusukebe/tap/rj

Binaries and go install are also available — see the repo.

Usage

rj https://skills.yusuke.run/skills.json

Output looks like:

{
  "code": 200,
  "header": { "content-type": "application/json" },
  "protocol": "HTTP/1.1",
  "status": "200 OK",
  "timing": { "dns_lookup": 0.48, "tcp_connection": 0.21, ... }
}

Pipe into jq to extract specific fields:

rj https://skills.yusuke.run/ | jq '.header'
rj https://skills.yusuke.run/ | jq '.code'

Common flags:

• -X POST — change method.
• -H 'Header: value' — add a request header.
• -u user:pass — Basic auth.
• --http3 / --http1.1 — pin HTTP version.

When to use

• Confirming response headers (content-type, cache-control, custom X-* headers).
• Verifying that an endpoint returns the expected status code.
• Checking how a request is routed (HTTP version, timing breakdown).
• Scripted health checks that need a structured view of the exchange.

For inspecting the response body, reach for curl (and pipe through jq if it's JSON). Use rj when the wrapper around the body is what matters.