mitmdump -s save.py

save.py

# save.py
# mitmdump -s save.py
"""
mitmproxy addon: save each response body to a separate file and print only
metadata + response headers (no bodies).

Filename format:
  <ct_main_sanitized>__<url_escaped>__<timestamp>.response

Example printed output block:
================================================================================
# captured: 2025-11-05T12:34:56-0600
# request: GET https://example.com/path?q=1
# filename: application_json__https%3A%2F%2Fexample.com%2Fpath%3Fq%3D1__170-
HTTP/1.1 200 OK
Content-Type: application/json; charset=utf-8
Content-Length: 123
...
================================================================================
"""

from mitmproxy import http
import os
import time
import sys
from urllib.parse import quote_plus

OUTDIR = "responses"
os.makedirs(OUTDIR, exist_ok=True)

def _ct_main_sanitized(ct_header: str) -> str:
    """
    Return a filesystem-safe main content-type string derived from Content-Type header.
    Examples:
      "application/json; charset=utf-8" -> "application_json"
      "text/html" -> "text_html"
      "" or None -> "unknown"
    """
    if not ct_header:
        return "unknown"
    main = ct_header.split(";", 1)[0].strip().lower()
    if not main:
        return "unknown"
    # replace problematic characters with underscores
    sanitized = main.replace("/", "_").replace("+", "_").replace(" ", "_")
    # also collapse any remaining characters that might be unsafe
    return "".join(ch for ch in sanitized if ch.isalnum() or ch in ("_", "-"))

def _make_filename(content_type_header: str, url: str, ts: int) -> str:
    ct_s = _ct_main_sanitized(content_type_header)
    # url-escape the absolute request url
    url_esc = quote_plus(url, safe="")
    # optional: limit very long escaped urls to keep filenames reasonable
    # (trim the middle if too long)
    MAX_URL_ESC_LEN = 512
    if len(url_esc) > MAX_URL_ESC_LEN:
        # keep head and tail
        head = url_esc[:200]
        tail = url_esc[-300:]
        url_esc = head + "__TRUNC__" + tail
    fname = f"{ct_s}__{url_esc}__{ts}.response"
    # make sure filename isn't absurdly long for filesystems
    if len(fname) > 255:
        fname = fname[:240] + "__cut__.response"
    return fname

def _print_headers_only(flow: http.HTTPFlow, filename: str) -> None:
    r = flow.response
    req = flow.request
    ts = time.strftime("%Y-%m-%dT%H:%M:%S%z", time.localtime())
    sep = "=" * 80

    # status line (include http_version when available)
    http_version = getattr(r, "http_version", "1.1")
    status_line = f"HTTP/{http_version} {r.status_code} {r.reason or ''}".strip()

    out_lines = [
        sep,
        f"# captured: {ts}",
        f"# request: {req.method} {req.url}",
        f"# filename: {filename}",
        status_line
    ]
    # response headers
    for k, v in r.headers.items():
        out_lines.append(f"{k}: {v}")

    out_lines.append(sep + "\n")
    sys.stdout.write("\n".join(out_lines))
    sys.stdout.flush()

def response(flow: http.HTTPFlow) -> None:
    """
    mitmproxy response hook:
      - save response.content (decompressed bytes) to file
      - print metadata and response headers only
    """
    try:
        req = flow.request
        r = flow.response

        # timestamp in seconds
        ts = int(time.time())

        # compute filename
        ct_header = r.headers.get("content-type", "")
        fname = _make_filename(ct_header, req.url, ts)
        path = os.path.join(OUTDIR, fname)

        # write raw uncompressed bytes (mitmproxy exposes decompressed bytes via .content)
        body_bytes = r.content if r.content is not None else b""
        try:
            with open(path, "wb") as fh:
                fh.write(body_bytes)
        except Exception as e:
            # still print headers but mark file write error
            _print_headers_only(flow, f"[error writing file: {e}]")
            return

        # print metadata + headers (no body)
        _print_headers_only(flow, fname)

    except Exception as exc:
        # never crash mitmdump; write error to stderr and continue
        sys.stderr.write(f"[save_headers_and_bodies.py error] {exc}\n")
        sys.stderr.flush()