**Note: These instructions won't help you get pass Cloudflare captchas, only through their firewall!!**
If you want to get proxied images from websites behind Cloudflare (or search engine data from the `Yep` search engine), you need to setup `curl-impersonate`. To do this, follow the install instructions listed here:
Only install the Firefox module. This is needed to fool Cloudflare into thinking we're making requests from the Firefox browser. Cloudflare is a piece of shit that checks TLS fingerprints and it can detect `curl` this way... By pretending to have Firefox's signatures, we can get pass their firewall.
Anyway, once you compiled all of this, you should be able to execute `curl_ff117 --version` and get an output similar to this:
Once you managed to do that, you still need to tell PHP to use the new library... This can be difficult to do depending of your distribution, but here is how I did it on Debian. First, you need to replace the `libcurl.so.4` file. You can use `find` to find where it is... Although it will return multiple matches.
```sh
$ sudo find / -name libcurl.so.4
/usr/lib/i386-linux-gnu/libcurl.so.4
/usr/lib/x86_64-linux-gnu/libcurl.so.4 # on debian, this is the one I must replace!
```
Once you found it, check if the library you compiled is present in `/usr/local/lib/libcurl-impersonate-ff.so`. If so, use the following commands:
Make sure you configure this right to optimize your search engine presence! Head over to `/robots.txt` and change the 4get.ca domain to your own domain.
To be listed on https://4get.ca/instances , you must contact *any* of the people in the server list and ask them to add you to their list of instances in their configuration. The instance list is distributed, and I don't have control over it.
If you see spammy entries in your instances list, simply remove the instance from your list that pushes the offending entries.
3. Go to the **main configuration file**. Then, find which website you want to setup a proxy for.
4. Modify the value `false` with `"myproxy"`, with quotes included and the semicolon at the end.
Done! The scraper you chose should now be using the rotating proxies. When asking for the next page of results, it will use the same proxy to avoid detection!
If you ever test out a `socks5` proxy locally on your machine and find out it works but doesn't on your server, try supplying the `socks5_hostname` protocol instead. Hopefully this tip can save you 3 hours of your life!