Fork of Blaine Cook's FakeWeb for fixes and new features
Switch branches/tags
Clone or download
Pull request Compare This branch is 2 commits ahead, 346 commits behind chrisk:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.



FakeWeb is a helper for faking web requests in Ruby. It works at a global level, without modifying code or writing extensive stubs.


This fork of Blaine Cook's original code has lots of fixes, stability improvements, and a few new features. To get it, install the latest gem directly from GitHub (currently

sudo gem install chrisk-fakeweb --source

Warning: this README goes with the latest development version, which has significant API changes. If you're using, please see the README for that release here:

Development Notes

We're currently considering how the API should change to add support for request bodies (see Known Issues below). Your input might be really helpful: see for a discussion of some different options. Thanks!


Start by requiring FakeWeb:

require 'rubygems'
require 'fake_web'

Registering basic string responses

FakeWeb.register_uri(:get, "", :string => "Hello World!")

=> "Hello World!"

=> FakeWeb is bypassed and the response from a real request is returned

Replaying a recorded response

page = `curl -is`
FakeWeb.register_uri(:get, "", :response => page)

# => Full response, including headers

Adding a custom status to the response

FakeWeb.register_uri(:get, "", :string => "Nothing to be found 'round here",
                                                  :status => ["404", "Not Found"])

Net::HTTP.start("") do |req|
  response = req.get("/")
  response.code     # => "404"
  response.message  # => "Not Found"
  response.body     # => "Nothing to be found 'round here"

Responding to any HTTP method

FakeWeb.register_uri(:any, "", :string => "response for any HTTP method")

If you use the :any symbol, the URI you specify will be completely stubbed out (regardless of the HTTP method of the request). This can be useful for RPC-like services, where the HTTP method isn't significant. (Older versions of FakeWeb always behaved like this, and didn't accept the first method argument above; this syntax is still supported, for backwards-compatibility, but it will probably be deprecated at some point.)

Rotating responses

You can optionally call FakeWeb.register_uri with an array of options hashes; these are used, in order, to respond to repeated requests. Once you run out of responses, further requests always receive the last response. (You can also send a response more than once before rotating, by specifying a :times option for that response.)

FakeWeb.register_uri(:delete, "",
                     [{:string => "Post 1 deleted.", :status => ["200", "OK"]},
                      {:string => "Post not found",  :status => ["404", "Not Found"]}])

Net::HTTP.start("") do |req|
  req.delete("/posts/1").body  # => "Post 1 deleted"
  req.delete("/posts/1").body  # => "Post not found"
  req.delete("/posts/1").body  # => "Post not found"

Clearing registered URIs

The FakeWeb registry is a singleton that lasts for the duration of your program, maintaining every fake response you register. If needed, you can clean out the registry and remove all registered URIs:


Blocking all real requests

When you're using FakeWeb to replace all of your requests, it's useful to catch when requests are made for unregistered URIs (unlike the default behavior, which is to pass those requests through to Net::HTTP as usual).

FakeWeb.allow_net_connect = false
=> raises FakeWeb::NetConnectNotAllowedError

FakeWeb.allow_net_connect = true
=> FakeWeb is bypassed and the response from a real request is returned

This is handy when you want to make sure your tests are self-contained, or you want to catch the scenario when a URI is changed in implementation code without a corresponding test change.

More info

FakeWeb lets you decouple your test environment from live services without modifying code or writing extensive stubs.

In addition to the conceptual advantage of having idempotent request behaviour, FakeWeb makes tests run faster than if they were made to remote (or even local) web servers. It also makes it possible to run tests without a network connection or in situations where the server is behind a firewall or has host-based access controls.

FakeWeb works with anything based on Net::HTTP–both higher-level wrappers, like OpenURI, as well as a ton of libraries for popular web services.

Known Issues

  • Request bodies are ignored, including PUT and POST parameters. If you need different responses for different request bodies, you need to request different URLs, and register different responses for each. (Query strings are fully supported, though.)


Copyright 2006-2007 Blaine Cook

Copyright 2008-2009 various contributors

FakeWeb is free software; you can redistribute it and/or modify it under the
terms of the GNU General Public License as published by the Free Software
Foundation; either version 2 of the License, or (at your option) any later

FakeWeb is distributed in the hope that it will be useful, but WITHOUT ANY
WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
FOR A PARTICULAR PURPOSE. See the GNU General Public License for more

You should have received a copy of the GNU General Public License along
with FakeWeb; if not, write to the Free Software Foundation, Inc., 51
Franklin St, Fifth Floor, Boston, MA 02110-1301 USA

See LICENSE.txt for the full terms.