A website usually consists of multiple objects (e.g. text file, a few png files etc.), I would like to know if there's a tool that can capture the individual requests/responses in different pcap files?
So for example if I browse to http://somewebsite.com , and some http://somewebsite.com consists of say {index.html, image1.png, somestylefile.css, image2.png}, the tool would capture the entire load of http://somewebsite.com but generate {index.html.pcap, image1.png.pcap, somestylefile.css.pcap, image2.png.pcap}
I don't know of any tool that can do this, or its possible using scapy or something similar?
A HTTP connection can have multiple requests inside the same TCP connection and browsers make heavy use of this HTTP keep alive. With HTTP pipelining the requests/responses don't even need to be fully separated in time, i.e. a client can send another request even though the response for the previous one is not there. And with HTTP/2 the data can also be interleaved, i.e. several responses transferred at the same time inside the same connection.
Insofar it is not always possible to capture the data as separate pcap file because they might not be separable at the packet level. But if you don't need the original packet boundaries it would be possible to create separate pcap files for each request which not necessarily reflect the original packets but which reflect the application layer, i.e. the response matching the request.
One tool which makes this is httpflow.pl which can extract HTTP/1.x requests/response pairs from an existing pcap (or sniff directly) and writes each request/response into a separate pcap file, as if it would have been a separate TCP connection. It can also clean up the data for easier analysis, i.e. unchunk and uncompress the HTTP body.