Skip to content
proton-cli
Menu
DocsCommandsFAQStats
Start here
InstallFirst commandsHow a command is builtManaging proton itself
Everyday use
Naming what to act onFilters and bulk changesDry runs and confirmationsOutput and exit codesScriptingAI agentsSettings, files and environment
When something goes wrong
TroubleshootingWhat it can't do
Account
Guideproton accountproton account keysproton account sessionsproton account security-logproton account breachesproton account profilesproton account settings
Mail
Guideproton mail messagesproton mail conversationsproton mail draftsproton mail mailing-listsproton mail protectedproton mail settings
Drive
Guideproton drive itemsproton drive trashproton drive photosproton drive computersproton drive sharingproton drive linksproton drive sharedproton drive invitationsproton drive settingsproton drive volumes
Calendar
Guideproton calendar eventsproton calendar busy-timesproton calendar booking-pagesproton calendar remindersproton calendar invitationsproton calendar settings
Pass
Guideproton pass itemsproton pass vaultsproton pass aliasesproton pass linksproton pass sharingproton pass sharedproton pass invitationsproton pass trashproton pass breachesproton pass settingsimport, export, generate
Contacts
Guideproton contactsproton contacts emailsproton contacts groupsproton contacts keys
Local index
Guideproton index
Raw API
Guideproton api
About
FAQSecurityAll commandsChangelogProton API reference

Stats

How often proton-cli is installed and updated, from the public counters every release channel leaves behind.

Every number here counts an acquisition: a release asset fetched, a page visited, a star given. None of it comes from the program. proton-cli has no telemetry, and installing it tells nobody anything - these are the counters GitHub, npm, Repology and the AUR publish about the project, recorded once a day so they have a history.

Loading…

What these numbers are not

A download is not a person. Release assets are fetched by mirrors, scanners and continuous integration as well as by people, and nothing here reports whether what was downloaded was ever run, or run twice.

Some channels leave no counter at all. Nix builds from source and serves from a shared binary cache, go install is answered by a global module proxy, and apt installs come from this project's own repository, which keeps no logs. Whatever those account for is missing from every figure above.

Clones count automation as readily as people, because a checkout in continuous integration is a full clone and runs on every push. That is why the figure here is unique cloners rather than clones.