Is there a way that I can save all HTTP and HTTPS browser communications (including request, response, full headers and body) to files on my computer?
The HTTPS communications must be saved decrypted.
Ideally I'd be looking for something like an extension for Firefox or Google Chrome.
It must work on Linux.
HAR is HTTP Archive, and is exactly what you want to save (it includes HTTPS unencrypted).
Luckily this is now a basic feature of all major browser (it was not when you asked the question). No more need of third party extensions.
In Firefox (v41+):
Other browser:
In recent version of all browser, you can save HAR easily:
HTTPFox is a Firefox extension that will do the job.
===EDIT===
Right click the request panel and click "Copy all rows". Then you can save it to what ever editor you like - Openoffice spreadsheet for example. Please see the image that I provided.
Error loading content (NS_ERROR_DOCUMENT_NOT_CACHED). But I need the response body of all requests, so HttpFox can't be a solution to this problem.
Fiddler is definitely something you should consider even if it is not an extension but a separate application.
I found this answer wanting to do this, and none of the answers here suited me because I needed to dump HTTPS response bodies, much like OP. So, for those who land here like I did, here's what I found.
I eventually landed on mitmproxy, which isn't a Firefox extension, but does accomplish the goal of logging all Firefox traffic. It's all Python, so I installed it with pip install mitmproxy - use a package manger if you like.
Started it up with mitmproxy for initial setup/debugging, and then used mitmdump to dump the stream once I had it working and refined.
I just set Firefox to use 127.0.0.1:8080 as the proxy server (8080 being mitmproxy's default port), and I saw all my Firefox traffic coming through mitmproxy. To enable HTTPS traffic, you have to accept the certificate that mitmproxy creates - with the proxy set up, just go to the "magic address" of http://mitm.it in Firefox, click the "Other" button, check the boxes to trust the cert, and hit OK.
To dump the unencrypted response body, I had to use a very simple inline script:
from libmproxy.model import decoded
def response(context, flow):
with decoded(flow.response): # automatically decode gzipped responses.
with open("body.txt","ab") as f:
f.write(flow.response.content)
Gist here - download it as save_response.py and use it with mitmdump -s save_response.py. Response bodies will start piling up in body.txt.
mitmproxy also has a number of useful filters that you can specify to mitmdump to grab just what you need.
httpFox, httpScoop (Mac only) and httpWatch (IE, Firefox, windows only) are all tools I use for doing this. Additionally you can also try Fiddler (as previosuly suggested) and also if yuo're using a Mac you could try Charles Proxy. My personal favourite is httpWatch (you can easily filter on headers, content, url etc), but this is expensive so may not be an option.
You might also want to look at the HAR project (HAR stands for HTTP Archive) and is a standards based approach to recording http transactions, headers, content etc. A number of the tools mentioned are able to export thee HAR files for viewing using he HAR Viewer (which you'll be familiar with if you've looked at the waterfall feature in Firebug). In fact, you can use Firebug to generate these files/logs for you as well using NetExport.
FWIW httpScoop is a good method for debugging/tracing HTTP traffic on a wireless network, I wrote an article about doing this some time ago: http://blog.adtools.co.uk/trace-debug-mobile-application-http-requests-using-macos/
Wireshark can be used to capture network packets, including things in the http protocol layer, and save them to your computer. It can also capture https information, but I doubt that you'd be able to configure it to decrypt anything.
