Skip to content
Create a gist now

Instantly share code, notes, and snippets.

Embed URL


Subversion checkout URL

You can clone with
Download ZIP
bulk-export run data from Garmin Connect
#!/usr/bin/env ruby
## disconnect
# ./disconnect.rb -u yourusername
# This is a command-line utility for the bulk-downloading of run data from
# the web application, which has lackluster export
# capabilities.
# Using this code is a matter of your own relationship with Garmin Connect
# and their TOS. I can't imagine this being very destructive to their service,
# and it's just filling in a hole in their existing service.
# It's built against Garmin Connect as of July 22, 2011. It's a scraper:
# thus if Garmin changes, this **will break**.
# This script requires all of the utilities on the line below: install them
# with rubygems
%w{rubygems json fileutils mechanize choice highline/import}.map{|x| require x}
Choice.options do
header ''
header 'Specific options:'
option :user, :required => true do
short '-u'
long '--user=USER'
desc ' username. Required'
option :dir do
short '-o'
long '--output-dir=OUTPUT'
desc 'the directory to save .tcx files'
default 'tcx'
password = ask("Enter your password: " ) { |q| q.echo = "*" }
def login(agent, user, password)
page = agent.get(LOGIN_PAGE)
login_form = page.form('login')
login_form['login:loginUsernameField'] = user
login_form['login:password'] = password
page = agent.submit(login_form, login_form.buttons.first)
raise "Login incorrect!" if page.title().match('Sign In')
def download_run(agent, id)
print "."
# This downloads TCX files: you can swap out the constant, or add
# more lines that download the different kinds of exports. I prefer TCX,
# because despite being a 'private standard,' it includes all data,
# including heart rate data.
agent.get(TCX_EXPORT % (id).to_i).save_as(File.join(Choice[:dir], "%d.tcx" % id))
def activities(agent)
j = agent.get(ACTIVITIES_SEARCH)
search = JSON.parse(j.content)
runs = search['results']['activities'].map {|r|
# Get each activity id to insert into the download URL
}.map {|id|
# Download a run.
download_run(agent, id)
agent =
# One needs to log in to get access to private runs. Mechanize will store
# the session data for the API call that cames next.
home_page = login(agent, Choice[:user], password)
FileUtils.mkdir_p(Choice[:dir]) if not[:dir])
puts "Downloading runs..."

If you get 400 => Net::HTTPBadRequest, it is possibly because Garmin's servers appear to be limiting activity queries to 100 requests at a time. Modify start and limit in ACTIVITIES_SEARCH and try again.


Works perfectly! Thanks a lot for this. just used it to export a couple of years of cycling data from Garmin and import it into Strava.


Works perfectly! Thanks a lot!

@pza --> thanks for your comment!


Thanks for great inspiration on writing Garmin Connect backup script.


yeah, thanks for this.


Ok. help me out here. Where and how do I copy and paste the above code? Thanks.


Thanks a lot for this, but.. I get Net::HTTPBadRequest even befgore I get one activity, and Garmin's "Activity Search Service 1.0" is deprecated: Is it over..?


@cloveras: try @gljeremy's fork:

@dano11 you'd copy & paste it into a file called disconnect.rb, and you'll need to gem install the dependencies on the top and run it with Ruby in a terminal.


Yes! Works perfectly, thanks a lot for the quick reply!


I note the TCX file downloaded does not contain the name or description associated with and activity on the garmin connect website, even though the file format supports that. :-( Can this script be extend to down loads as well?


Yes, thanks a lot for sharing. I also experienced the 400 => Net::HTTPBadRequest problem. Instead of limiting ACTIVITIES_SEARCH to 100, I limited it to 1 and changed the activities function to look like the following:
def activities(agent)
$activityIdx = 0
$numActivities = 624
while $activityIdx <= $numActivities do
puts("Downloading activity = #$activityIdx")
as = ACTIVITIES_SEARCH % $activityIdx
print "\tas: ", as, "\n"
j = agent.get(as)
search = JSON.parse(j.content)
runs = search['results']['activities'].map {|r|
# Get each activity id to insert into the download URL
}.map {|id|
# Download an activity.
download_run(agent, id)
$activityIdx += 1

In my case I wanted all of my history, 624.

Thanks a lot for sharing @tmcw!


Garmin Connect UI has changed and this script no longer works for me.


If you're interested in a Python solution, I have one that does pretty much the same thing (in fact, this script was inspiration for it) and also contains a fix for the new Garmin Connect UI (basically just needs a redirect after login):

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Something went wrong with that request. Please try again.