Critical LaTeX Rendering Bug and Native App Regressions

As an active Memo AI Pro subscriber, I am writing to report several critical bugs and significant feature regressions encountered after switching from the Web App to the official App Store release (tested on both iPhone and iPad). These issues severely hinder daily usability.

1. LaTeX Rendering is Broken (Critical Usability Issue)

• Problem: Math formulas fail to compile and render properly, displaying raw LaTeX syntax instead (e.g., $\Delta=\begin{vmatrix}...$, \frac{...}{...}).

• Impact: For STEM and mathematics study, this makes the generated output illegible and effectively unusable. (Screenshots attached for reference)

2. Regressions in the Native App vs. Web App

• Missing "Thinking Effort" & Model Transparency:

  • In the Web App, users could configure the Memo Agent's thinking effort (e.g., Pro mode). In the mobile app, this option is completely gone.

  • Model selection lacks transparency. I only discovered the agent was running on deepseek/deepseek-v4-flash-0731 after checking the background "Usage" logs. Pro users need explicit control over model tiers and reasoning effort.

• No In-App Feedback Channel:

  • The native app completely omits an in-app feedback or bug-reporting option, forcing users to seek external channels to report urgent bugs.

• Profile Update Sync Failure:

  • Editing profile details prompts a success toast ("Your profile has been updated"), but restarting or re-entering the app reverts the profile back to its previous state.

  • This bug is persistent and was consistently reproduced across multiple attempts on both iPhone and iPad.

Given that these features worked reliably on the Web platform, the mobile app currently feels like a major downgrade for paying Pro users. Please prioritize restoring proper LaTeX parsing and feature parity for model controls and basic account synchronization.