VCL based device detection for Varnish Cache.
Branch: master
Clone or download
lkarsten Convert (most) scripts to Python 3.
Missing: the development web server. I don't think that sees much use,
we can convert it to http.server when/if someone ever needs it on a box
without python2.
Latest commit 2bcd548 Jul 31, 2018
Type Name Latest commit message Commit time
Failed to load latest commit information.
backend Convert (most) scripts to Python 3. Jul 31, 2018
.gitignore Big rework of example VCL testing. Dec 19, 2014
.travis.yml Point Travis to Packagecloud Jun 25, 2018
INSTALL.rst Use elsif everywhere for consistency Jun 23, 2015
LICENSE Add a license file. Solves #5 Mar 18, 2013
Makefile Wildcard expansion time in make is complicated. Dec 19, 2014
controlset.txt Fix space-vs-tab issue introduced in 6210e0b. Jul 31, 2018


Device detection in Varnish


The goal of this VCL set is to provide a simple & easy way of getting device detection going in Varnish Cache. (

Specific problems we want to solve:

  1. Detect the most common and easily detected (mobile/tablet) platforms.
  2. Provide example VCL for using devicedetect.vcl with your existing VCL.

These are explicit not goals for this project:

  • Create a comprehensive set of capabilities per client. We don't want to spend a lot of effort on maintenance.
  • Be perfect. It usually works, but if you need guarantees you should consider the commercial offerings instead.

This project is maintained and updated by the community. If you see any false positives or missing strings, fork the git repository and send a pull request.


You need a recent Varnish release. It may function with some adjustments on previous versions, but you are on your own.

It is worth noting that there is no compilation/linking required. This is VCL code only.

Your backends needs to be able to do produce different content based on what type of device this is. This can be signalled to the backend through the URL, with an extra HTTP header or with a cookie. See the INSTALL.rst file for examples.

Use cases

The following uses are envisioned for this VCL:


See the INSTALL.rst file for details.

Similar efforts

These similar efforts for User-Agent insights are known to us:


This project lives on Github:

Feature requests, bug reports and such can be added to the Github issue tracker.

This code is currently maintained by Christopher Slowe <>.