---
title: How to download rtmp streams?
---

> Real Time Messaging Protocol (RTMP) was initially a proprietary protocol
> developed by Macromedia for streaming audio, video and data over the Internet,
> between a Flash player and a server. Macromedia is now owned by Adobe, which has
> released an incomplete version of the specification of the protocol for public
> use.

Many media sites (N24, zdf, ard, ...) use rtmp streams to prevent people from
downloading the videos. Well, there is a little command line tool which allows
downloading rtmp streams. It is called
[rtmpdump](https://www.archlinux.org/packages/?name=rtmpdump) and is available in
the arch linux repositories.

To use rtmpdump you have to extract the required parameters from the html source
code. For N24 it looks like [this](/2014-09-22-n24-source.txt) (it is *really* ugly!).


To use rtmpdump we need to set the `-r`, `-W` and `-y` parameters; with `-o` we
can specify a filename. An explanation of these parameters could be found with
`rtmpdump -h`. An appropriate rtmpdump call for the html code above would look
like the following:

```
$ rtmpdump \
  -r "rtmp://pssimn24livefs.fplive.net:1935/pssimn24/" \  # <1>
  -W "http://www.n24.de/_swf/HomePlayer.swf?cachingVersion=2.66" \  # <2>
  -y "mp4:cm2013/cmp/e214/89ed/ac91/30c0/0a1db4af7089b0c05161ea2dcca58546_1000.mp4" \  # <3>
  -o "myvideo.mp4"
```

As you can see I have extracted these values from the
[source code](http://pastebin.com/raw.php?i=brF5DvUL) of the website. For n24 it
is:

1. `_n24VideoCfg.flash.videoFlashconnectionUrl`
2. `_n24VideoCfg.flash.SWFUrl` (prepend the baseurl!)
3. `_n24VideoCfg.flash.videoFlashSource`

For simplification I have written a little python wrapper which extracts these
parameters from a given URL and builds the rtmpdump command line string:

```
import re
import sys
import subprocess
import urllib.request
from urllib.parse import urlparse


url = sys.argv[1]
url_parsed = urlparse(url)
response = urllib.request.urlopen(url)
html = response.read().decode()

regex = re.compile(
    r'<title.*?>(?P<title>.+?)</title>|'
    r'(?:_n24VideoCfg.flash.videoFlashconnectionUrl = )["\'](?P<param_r>.*)["\'];|'
    r'(?:_n24VideoCfg.flash.SWFUrl = )["\'](?P<param_W>.*)["\'];|'
    r'(?:_n24VideoCfg.flash.videoFlashSource = )["\'](?P<param_y>.*)["\'];|'
)

title = ''
rtmpdump_params = {}

for match in regex.finditer(html):
    typ = match.lastgroup

    if typ == 'title':
        title = match.group('title')
    elif typ == 'param_r':
        rtmpdump_params['-r'] = match.group('param_r')
    elif typ == 'param_W':
        rtmpdump_params['-W'] = match.group('param_W')
    elif typ == 'param_y':
        rtmpdump_params['-y'] = match.group('param_y')

subprocess.call([
    'rtmpdump',
    '-r', rtmpdump_params['-r'],
    '-W', url_parsed[0]  + '://' + url_parsed[1] + rtmpdump_params['-W'],
    '-y', rtmpdump_params['-y'],
    '-o', title + '.mp4'
])
```

UPDATE 2020-08-24: Just use [`youtube-dl`](https://ytdl-org.github.io/youtube-dl/). It does everything automagically and is maintained.