cURL / HTTP request tool

The cURL / HTTP request tool reads a cURL command — the one from a browser's Copy as cURL, an API's docs, a colleague's message — into a form you can edit: method, URL, headers, body, auth. From the form it writes the request four ways, each ready to copy: cURL tidied up, fetch, axios and Python requests; the request itself is never sent.

The method, headers, cookies, user agent and referer; every form of data — -d, --data-raw, --data-binary, --data-urlencode, --json — and multipart forms with -F; basic auth with -u; -G, -L, -k and --compressed; the URL with or without --url. Single and double quotes, backslash escapes, lines continued with a backslash, attached values like -XPOST, --long=value and bundles like -sSL are all read. Options that do not change the request — silence, verbosity, output files, timeouts, proxies — are read and ignored.

Never. There is no Send button and no network call; the page turns one text into another, so a token or a cookie in the command goes nowhere. To run the request, copy the output into your terminal, browser console, script or notebook.

From the browser: open DevTools, the Network tab, right-click the request and choose Copy as cURL. Everything the page sent — headers, cookies, body — comes with it, which is why it is worth reading before you share it.

When the body is JSON and the content type says so, fetch writes it as JSON.stringify({...}), axios as a data object and Python as json={...}. Other bodies are passed as strings, and -u becomes an Authorization header in fetch, the auth option in axios and auth=(user, password) in Python, with the credentials as you typed them.

The form comes out as FormData in fetch and axios and as data= and files= in Python. A file field (@path) cannot travel through a text: JavaScript gets a placeholder that expects a file input, and Python gets open(path, "rb") with the path from the command.