the Chromium logo

The Chromium Projects

Have you ever hit a regression bug like this: "In chromium 85.0.4183.121, things were broken. Back in chromium 86.0.4240.193, it was fine."? A good way to attack bugs like this – where it's unclear what change could have caused the regression, but where you have a reliable repro – is to bisect.

tools/ automates downloading builds of Chrome across a regression range, conducting a binary search for the problematic change.

If you don't have a chromium checkout, you can fetch just this script with the commands below:


curl -s --basic -n "" | base64 -d >


curl -s --basic -n "" | base64 -D >

Windows (no curl, base64) python3

python3 -c "import base64; import urllib.request; print(str(base64.b64decode(urllib.request.urlopen(\"\").read()),'utf-8'))" >

Run it like this:

python tools/ -a platform -g good-revision -b bad-revision -- flags-for-chrome

For example,

python tools/ -a mac -g 782793 -b 800218 --use-local-cache --verify-range -- --no-first-run --user-data-dir=/tmp

Also, you can specify either end of the bisect range using version numbers.

python tools/ -a mac -g 85.0.4183.121 -b 86.0.4240.193 --use-local-cache --verify-range

The two range specifications above are equivalent. Note that in all cases the bisect is being done from trunk builds so merges to a release branch will not be bisected.

Note: script now requires Python 3 to run for Windows. Use python3 on Windows instead, in order to run the examples above.

Valid archive types (the -a parameter) are mac, mac64, win, win64, linux (not supported for builds after March 2016), linux64, linux-arm, and chromeos.

You can also use the -p option to specify a profile. If no -p or --user-data-dir option is specified, a new profile will be created in a temporary directory each time you are asked to try a build. If you specify a profile folder, point to the directory that's a parent of Default/.

The script will download a build in the revision range and execute it. You must then manually check if the bug still repros. Quit Chromium, and the script will ask you if the bug reproduced or not. It will use your answer to drive a binary search, and after just a few steps it will tell you "this regression happened somewhere between revisions 793241 and 793248". From that list, it's usually easy to spot the offending CL. If you're adding the range as a comment to a bug, please always paste the output from, as this includes links to the chromium changes in the regression range.

View code changes in revision range with this useful URL (replacing SUCCESS_REV and FAILURE_REV with the range start and end):

Notes: The default option is snapshot(Chromium build). There are also release build(-r) and official build(-o) available for Googlers. Googlers should use official build(-o) whenever possible to bisect to a single commit. Please refer to go/chrome-bisect for more information.

Getting an initial revision range

If you have two Chrome binaries, one which doesn't work, one which does, you can find their revision numbers as follows.

First, visit the [chrome://version](javascript:void(0);) page and copy the version number (for example 85.0.4183.121). Then, use this version number as the parameter to -g or -b. Alternately you can use stable-release milestone numbers (M85, M86) or look in git commit messages for "Cr-Commit-Position: refs/heads/main@{%825204}".

Verifying the range

If your revision range is incorrect, or if something about your environment interferes with your reproduction of the bug, you will not get useful results from If you would prefer to know this as soon as possible, rather than after downloading and checking O(log n) builds, pass the --verify-range option to This will check the first and last builds in the range before starting the bisect.

Field trials

The bisect script uses non-Chrome branded builds and therefore uses the field trial testing config. This means that behavior may differ from what you see in Chrome. Consider using the disable-field-trial-config command line switch if this matters to you.

To find out which variation causes a regression, you can use

API Keys and Chrome OS builds

Without API keys, Chrome OS won't allow you to log in as a specific user. To run a chromeos bisect on your Linux desktop, add the following variables to your environment (e.g., via .bashrc):




See for more info about API keys.