Anyone else's citation styles randomly resetting? It's driving me nuts.

I've been conducting a detailed security and compliance review of SciSpace for my organization's approved SaaS tool list, with a particular focus on its data persistence, configuration management, and session handling. During this extended evaluation period, I've encountered a persistent and operationally disruptive bug: the citation style selection within the manuscript editor does not maintain state consistently.

The issue manifests as follows: after setting a specific citation style (e.g., APA 7th, IEEE, Nature), the application will, without warning or apparent trigger, revert to a default style (often appears to be MLA). This occurs across multiple browser sessions and persists even after explicit saves and exports. My testing methodology has attempted to isolate variables:

* **Browser & Cache:** Reproduced in clean, isolated sessions on Chrome, Firefox, and Edge with all extensions disabled. Clearing local storage, session storage, and cookies does not prevent recurrence, but does reset the faulty state, suggesting the problem is server-side profile or document-state related.
* **Document Actions:** The reset seems uncorrelated to any single user action I can pinpoint. It has occurred during idle periods, after using the "Save" function, after exporting a draft, and upon simply refreshing the page.
* **Data Handling Concern:** From a data integrity perspective, this is significant. If a core piece of document configuration (citation style dictates final output format) is not being reliably persisted, it calls into question the robustness of their user data model and the transactionality of their save operations. For a tool handling academic and potentially proprietary research, this is a serious workflow and compliance pitfall.

My initial investigation suggests this is not a local environment issue. I've reviewed browser developer console logs and network activity, and there are no consistent errors thrown when the reset happens, which is even more troubling—it appears to be a silent failure in their state synchronization.

I am seeking to determine if this is a widespread systemic issue or something isolated to a specific document type or account configuration. Has anyone else performed a methodical analysis of this behavior?

* What citation style were you using when it reset?
* Can you correlate the reset with any specific API call or UI interaction? (Check your browser's Network tab for calls to `graphql` or `rest` endpoints around the time of the reset).
* Does the issue persist in a completely new, blank document, or only in documents imported from certain file formats?

Gathering this data is crucial not only for a workaround but to assess whether this represents a deeper architectural flaw in their application's handling of user preferences. For my SOC2 compliance review, consistent configuration management is a key control point, and this behavior would be a major finding.

- RayS