A web proxy for evading corporate or government filters, similar to CGIproxy / PHProxy / Glype but written in node.js. All data is processed and relayed to the client on the fly without unnecessary buffering.
Any website that the proxy can access can now be reached by the proxy's users.
The script uses "pretty" urls which, besides looking pretty, allow links with relative paths
to just work without modification. (E.g. <a href="path/to/file2.html"></a>
)
In addition to this, links that are relative to the root (E.g. <a href="/path/to/file2.html"></a>
)
can be handled without modification by checking the referrer and 307 redirecting them to the proper
location in the referring site. (Although the proxy does attempt to rewrite these links to avoid the redirect.)
Cookies are currently storred in the visitor's session on the server rather than being sent to the visitor's browser to avoid having a large number of (possibly conflicting) browser cookies once they have browsed several sites through the proxy.
Requires node.js >= 0.8 (0.10 is recommended) and Redis for session storage.
Then download node-unblocker, cd into the directory,
and run npm rebuild
. Optionally edit
config.js then run npm start
to start the server. It should spawn a new instance for each CPU
core you have.
(Note: running node app.js
will not work. The server code is in the Gatling package, which the npm start
command calls automatically.)
This project should be runnable on a free Heroku instance without modification - see http://node-unblocker.herokuapp.com/proxy for an example. You will want to run the following commands:
heroku addons:add redistogo
heroku config:add SECRET=<TYPE SOMETHING SECRET AND/OR RANDOM HERE>
heroku addons:add piggyback_ssl
heroku addons:add newrelic:stark
This sets up a free redis cache instance, secures your cookies, and adds https support, and sets up free monitoring with newrelic (optional). You may also want to run this to enable usage tracking via Google Analytics:
heroku config:add GA_ID=[your Google Analytics ID, ex: UA-12345-78]
- Write more tests: character encoding, compression, end-to-end tests in real browsers
- Consider gzipping all appropriate responses (anything text-like and more than a few kb)
- Mini-url form
- Allow for removal of scripts (both script tags and on*= handlers)
- Figure out how to make the blocklist work with heroku without shipping a default one with the source.
This project is released under the terms of the GNU GPL version 3
- Replaced server.js with Gatling
- Removed memwatch
- Updated design to be mobile-friendly
- Tweaked Redis client and blocklist to not keep server open after unit tests
- Seperated app and server more cleanly
- Additional JSHint checks
- Added memwatch
- Replaced built-in monitoring code with (optional) New Relic support
- Split proxying and server code into two files.
- Fixed a bug when attempting to parse cookies on invalid urls
- Added JSHint to the test suite
- JSBeautify'd code
- Moved static content and code to it's own directory and file
- Added a test for static content
- Added unit tests for url prefixing on streams that get split in various locations
- Fixed bugs these tests revealed
- Fixed bug with links pointing to / not getting rewritten
- Added backpressure support to streams
- Unit tests for Google Analytics
- Google Analytics bug fix
- Set up Continous Deployment
- Default proxied traffic to SSL if url is nodeunblocker.com
- Updated to Node.js v0.10-style streams
- Split encoding, url prefixing, ROBOTS meta tag, and Google Analytics into individual files (and streams)
- Unit tests for UrlPrefixStream.
- Added a performance test
- Increased the HTTP Agent's maximum number of Open Connections - issue nfriedly#17
- Added tests
- Fixed a few bugs with creating proxied links.
- Added support for more charsets via Iconv. (Issues #10 & #11) ** This may have broken compatibility with Windows, more investigation to come. https://github.com/nfriedly/node-unblocker/zipball/v0.7.1 is pure JS and known to be Windows-compatible.
- Added GA tracking and and noindex/nofollow meta tags to proxied pages
- Improved status page to show cluster-wide statistics (Issue #4)
- Fixed issue #7 to better track concurrent requests
- Added support for node.js 0.6's native clustering
- Removed simple-session library and replaced it with connect's session library backed by a redis store
- Reworked fileserver to serve index.html from memory and use compression when avaliable
- Added some windows support (although it doesn't bind to localhost)
- Fixed issue #2 for relative path bug when the domain name didn't have a / following it
- Removed compress library dependency in favor of the native zlib library that shipped in node 0.6
- Several small tweaks to support running on Heroku servers
- Added keyword and domain blocklists
- Pulled out configuration into a separate file
- Set up live demo at nodeunblocker.com
- Added "military" theme
- Added support for remote HTTPS servers.
- Created a simple-session library. (The ones I tried were all tied to bigger projects and/or didn't work well)
- Added basic cookie support via sessions.
- Urls that are relative to the root of the site are now processed in both html and css.
- Now only buffers last few characters if a chunk appears to end in the middle of a url.
- Added redirect support
- Added gzip support
- improved filters
- Initial release; basic passthrough and url-fixing functionality