Cherry Studio, a popular open-source AI client, has come under scrutiny after users found that its privacy setting did not fully stop telemetry. According to GitHub user Yuerchu, the app continued to send data to analytics.cherry-ai.com even when the option labeled “Send anonymous error reports and usage statistics” was turned off.
The reported data includes a unique device ID, system information, CPU architecture, and app version. Based on the disclosed details, the issue does not affect chat content or other sensitive conversation data. Instead, the concern centers on device-level and application-level analytics being transmitted despite user preferences.
A March Code Change Triggered the Behavior
The report links the problem to a code change made by maintainer kangfenmao on March 22. That change reportedly caused app launches and update checks to bypass the privacy toggle, meaning some analytics events were still sent even after users opted out.
The behavior persisted across several versions, suggesting it was not an isolated release bug. For an open-source AI application, where transparency and user control are key expectations, a mismatch between settings and actual data handling can quickly become a trust issue.
Fix Deployed, but Some Users Should Recheck Settings
A fix has now been implemented to make the privacy toggle apply consistently across all event types. At the same time, developers uncovered an older issue affecting upgrades from versions released before February 2025. In those cases, the upgrade script could re-enable the privacy option, which means some long-time users may need to manually disable it again.
Overall, the incident appears to be less about sensitive content exposure and more about whether user choices were properly respected. Users relying on Cherry Studio or similar AI tools may want to verify their privacy settings after updating and confirm that telemetry behavior matches what the interface promises.