How do I get the user agent with Flask?

flask.request.user_agent.string

from flask import request
request.headers.get('User-Agent')

You can also use the request.user_agent object which contains the following attributes which are created based on the useragent string:

• platform (windows, linux, macos, etc.)
• browser (chrome, firefox, msie, etc.)
• version
• language
• string (== request.headers.get('User-Agent'))

Note: As of werkzeug 2.0, the parsed data of request.user_agent has been deprecated; if you want to keep getting details you need to use a custom UserAgent implementation and set it as user_agent_class on a custom Request subclass, which is set as request_class on the Flask instance (or a subclass).

Here's an example implementation that uses ua-parser:

from ua_parser import user_agent_parser
from werkzeug.user_agent import UserAgent
from werkzeug.utils import cached_property




class ParsedUserAgent(UserAgent):
@cached_property
def _details(self):
return user_agent_parser.Parse(self.string)


@property
def platform(self):
return self._details['os']['family']


@property
def browser(self):
return self._details['user_agent']['family']


@property
def version(self):
return '.'.join(
part
for key in ('major', 'minor', 'patch')
if (part := self._details['user_agent'][key]) is not None
)

If you use

request.headers.get('User-Agent')

you may get: Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/45.0.2454.101 Safari/537.36

If you use

request.user_agent

you may get like this:

• user_agent.platform: windows
• user_agent.browser: chrome
• user_agent.version: 45.0.2454.101
• user_agent.language: None
• user_agent.string: Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/45.0.2454.101 Safari/537.36

UA usually does not contain language. If you want to get the language set in browser, you may use

request.accept_languages

It'll give you list of languages. E.g.

LanguageAccept([('en-US', 1), ('en', 0.5)])

To access the first value, you may use

request.accept_languages[0][0]

which will result in string

'en-US'

Detailed information about 'accept_language" header: https://www.w3.org/International/questions/qa-lang-priorities

The question begs for a lot more information. This library seems to fit the bill of collecting a lot of information out of flask, and has example calls to getting this information out of the application context.

https://pythonhosted.org/Flask-Track-Usage/

Usage gets stored in this format:

[
{
'url': str,
'user_agent': {
'browser': str,
'language': str,
'platform': str,
'version': str,
},
'blueprint': str,
'view_args': dict or None
'status': int,
'remote_addr': str,
'xforwardedfor': str,
'authorization': bool
'ip_info': str or None,
'path': str,
'speed': float,
'date': datetime,
},
{
....
}
]

Here is one of the places in the library where the data is collected:

https://github.com/ashcrow/flask-track-usage/blob/master/src/flask_track_usage/init.py around line 158

    data = {
'url': ctx.request.url,
'user_agent': ctx.request.user_agent,
'server_name': ctx.app.name,
'blueprint': ctx.request.blueprint,
'view_args': ctx.request.view_args,
'status': response.status_code,
'remote_addr': ctx.request.remote_addr,
'xforwardedfor': ctx.request.headers.get(
'X-Forwarded-For', None),
'authorization': bool(ctx.request.authorization),
'ip_info': None,
'path': ctx.request.path,
'speed': float(speed),
'date': int(time.mktime(current_time.timetuple())),
'content_length': response.content_length,
'request': "{} {} {}".format(
ctx.request.method,
ctx.request.url,
ctx.request.environ.get('SERVER_PROTOCOL')
),
'url_args': dict(
[(k, ctx.request.args[k]) for k in ctx.request.args]
),
'username': None,
'track_var': g.track_var
}