UserAgent parser comparison

We took 6136 user agents
from 5 test suites
and analyzed them with 9 providers.
Here you can see the results!
Provider Results Browser Rendering engine Operating system Device Model Brand Type Is mobile Is bot info_outline Parse time info_outline Actions
BrowscapPhp
5536
5171
4978
4975
5169
5169
1388
5169
4269
365
0.0475 Details
DonatjUAParser
v0.5.0
5857
5857
0
0
0
0
0
0
0
0
0.00067 Details
PiwikDeviceDetector
3.4.2
6076
5578
5074
5402
5690
4709
4781
5404
4577
316
0.01312 Details
SinergiBrowserDetector
5.1.0
5685
5378
0
4938
3807
482
0
0
3470
207
0.00117 Details
UAParser
v3.4.4
5980
5543
0
5212
4956
4899
4945
0
0
254
0.00831 Details
WhichBrowser
v1.0.1045
5775
5053
5045
4849
5700
4904
4749
5700
4407
75
0.07551 Details
Woothee
v1.2.0
5568
4956
0
0
5162
0
0
5162
3856
234
0.0006 Details
Wurfl
for API 1.6.3, db.scientiamobile.com - 2015-10-16 11:07:21
5762
4623
0
4847
5629
5092
5092
5629
4448
133
0.20953 Details
YzalisUAParser
0.2.0
4977
4758
4568
4036
474
367
474
373
373
0
0.00249 Details

Not enough?

You can go to the details of each provider, or see here the results of all analyzed user agents

View all results

Source of user agents

The user agents were taken mainly out of the testsuites of the providers above. Thanks to all who provided them!
Name Number of user agents
donatj/phpuseragentparser 147
piwik/device-detector 3909
whichbrowser/testrunner 1770
woothee/woothee-testset 229
yzalis/ua-parser 81

More informations

The primary goal of this project is simple. I wanted to know which user agent parser is the most accurate in each part - device detection, bot detection and so on...

The secondary goal is to provide a source for all user agent parsers to improve their detection based on this results.

You can also improve this further, by suggesting ideads at ThaDafinser/UserAgentParserComparison

The comparison is based on the abstraction by ThaDafinser/UserAgentParser
Comparison created 2015-11-20 10:09:25 | by ThaDafinser