mitmproxy/doc-src/features/responsestreaming.html

By default, mitmproxy will read the entire response, perform any indicated
manipulations on it and then send the (possibly modified) response back to
the client.  In some cases this is undesirable and you may wish to "stream"
the reponse back to the client.  When streaming is enabled, the response is
not buffered but is instead sent directly back to the client.  (If HTTP
chunked transfer encoding is enabled, the response will be streamed
back one chunk at a time.) This is especially useful for large binary files,
which are often not what you are trying to inspect, and while buffering
cause browser slows.

Streaming can be enabled on the command line for all responses which are
greater than a certain size.  Note that the SIZE argument below can accept
the usual prefixes (m, k, etc.)

<table class="table">
    <tbody>
        <tr>
            <th width="20%">command-line</th>
            <td>
                <ul>
                    <li>--stream SIZE</li>
                </ul>
            </td>
        </tr>
    </tbody>
</table>

<div class="page-header">
    <h1>Customizing Response Streaming</h1>
</div>

You can also use an inline script hook to write code to customize exactly
which responses are streamed.

The basic concept is simple:

$!example("examples/stream.py")!$

See [inline scripts](@!urlTo("scripting/inlinescripts.html")!@) for more
info on how to make and use inline scripts in general.

<div class="page-header">
    <h1>Things to Know</h1>
</div>

When response streaming is enabled, streamed response content will not be
recorded with the -w parameter.

Portions of the code which would have otherwise performed changes
on the response body will instead see an empty response body
and any attempts to modify it will be ignored.