Research Projects: conduct data-driven research with Simons before building any strategy. Findings are stored immutably with verifiable integrity, and every hypothesis you test strengthens the statistical rigor of the strategies you deploy. Full AI model catalog: choose from the latest models including Claude Fable 5, Opus 5, Gemini 3.1 Pro, Grok 4.1, Kimi K3, and more. Response effort control: choose Fast, Balanced, Thorough, or Max reasoning depth, synced across web and mobile. Model and effort are now adjustable directly from the Simons chat.

Security and reliability improvements. This update strengthens protection for your account and data and includes stability fixes and performance improvements.

Bug Fixes & Improvements - Fixed character encoding issue where Simons AI displayed garbled characters (â) instead of proper em dashes and special characters in chat responses - Fixed strategy creation failing silently — the app now properly awaits the save operation and displays error messages if creation fails - Reduced swipe gesture sensitivity on strategy and position cards to prevent accidental triggers while scrolling - Fixed exit strategy orders for short positions (now correctly sends "buy to cover" instead of always "sell"), added price validation, and prevented API calls in demo mode - Wired up real speech-to-text dictation in the strategy creation sheet (was previously a non-functional stub) - Fixed push notifications not registering — device tokens are now retried after login if the initial registration fires before authentication completes - Fixed Simons Notes displaying raw markdown links like [NVDA](https://data-lab?...) — symbols now render as clean bold text - Fixed notification categorization where Simons Notes received in the background were incorrectly filed under Strategy Alerts - Fixed exit strategy sheet using a fresh API service instance instead of the shared one, which could send orders to the wrong account - Fixed Simons AI only passing the first strategy type when creating a strategy from chat suggestions - Corrected tip text that told users to "swipe left" for exit strategies when the actual gesture is swipe right - Removed unused code and dead UI state for improved stability