Version 1.9 introduces a direct-but-safe application policy across Apple Intelligence, catalog models and imported GGUF models. Responses are now checked before they are shown, saved, spoken or exported. The model library is reorganized by objective size: Fast (1–2B), Balanced (3B) and Advanced (7–8B). Existing downloads, imports and conversations remain available. Model descriptions now focus on speed, memory, reasoning, writing, coding, attachments and device compatibility. This update also refreshes onboarding, personas, paywall copy and App Store positioning around clear answers without unnecessary fluff.

Fixed a bug where the daily message limit could stop resetting, leaving some free users unable to send new messages. The counter now resets correctly at midnight, and affected devices recover automatically on first launch of this update.

We've made improvements and fixed bugs to make the app better for you.

Fixed an issue where Pro subscribers could incorrectly see the daily message limit. If you purchased a subscription, your unlimited messages now work as expected.

What's new in this version: • Paste Hugging Face model or GGUF file links and import compatible local models • Added safer model import details, including revision, file size, RAM guidance, and template fallback • Create, edit, duplicate, and delete custom personas • Improved chat history tools, summaries, export, and copy actions • Clearer Pro restore status and purchase recovery messaging

What's new in this version: • Added in-app Hugging Face GGUF import • Added Siri shortcut suggestions in Settings • Unified Prompt Library with custom prompts, delete, and restore built-ins • Added Pro chat export and copy tools from chat and history • Improved voice setup, prompt tools, and the chat composer