Now that Youtarr has a large base of actual users, I'd like better visibility into what most installs look like - which OS, which Docker variant, bind mount vs named volume, etc. The database corruption risk in #588 is a recent example: I have no way to estimate how many users are at risk.
I plan to add optional telemetry to fix that.
Privacy considerations
- Opt-in only. Default is always off. I'll never auto-enable it.
- No identifying information, ever. No usernames, no IP addresses, no API keys, no Plex tokens, no channel names, no video titles, no file paths, no email addresses, no webhook URLs.
- Self-hosted. The data goes to a server I run at my house. No third party gets it.
- You can see exactly what gets sent. The consent screen previews your real payload, and you can view it any time from settings.
- Easy to turn off. One toggle in settings.
What gets collected
App version, OS, architecture, WSL/Docker Desktop detection, database setup (bind mount vs named volume vs external), whether auth is enabled, which notification services are configured (just names like "discord", never URLs), and rough counts of channels and videos.
Anything that could identify a user or their network is excluded. The full list will be documented before this ships.
Now that Youtarr has a large base of actual users, I'd like better visibility into what most installs look like - which OS, which Docker variant, bind mount vs named volume, etc. The database corruption risk in #588 is a recent example: I have no way to estimate how many users are at risk.
I plan to add optional telemetry to fix that.
Privacy considerations
What gets collected
App version, OS, architecture, WSL/Docker Desktop detection, database setup (bind mount vs named volume vs external), whether auth is enabled, which notification services are configured (just names like "discord", never URLs), and rough counts of channels and videos.
Anything that could identify a user or their network is excluded. The full list will be documented before this ships.