
Open the app and choose an AI model.

Start chatting by asking questions or beginning a conversation.

Get fast replies generated directly on your device, without an internet.
No internet required
Answers in seconds
Data stays on your device
No cloud storage needed
No third-party access to your conversations
Choose from a variety of powerful AI models optimized for your iPhone. Each model offers different capabilities and performance characteristics.
8-bit (MLX)
Sharper version of the 1B model; still small and generally safe on iPhones.
4-bit (MLX)
Balanced Llama 3.2 chat model; good quality but heavy enough to crash on some iPhones.
4-bit (MLX)
Good mix of quality and speed; lighter than 7B models and safe for most iPhones.
3-bit (MLX)
Higher-quality 7B chat model; heavy and may crash on some iPhones during long chats.
4-bit (MLX)
Top Qwen2.5 quality; very heavy, can crash if the phone is hot or many apps are open.
4-bit (MLX)
Ultra-small and super fast; answers are simpler and less detailed than bigger models.
4-bit (MLX)
Base (non-instruct) model; best with strong custom system prompts and advanced setups.
6-bit (MLX)
Same as 4B, but with more precision.
3-bit (MLX)
Large 8B base model.
8-bit (MLX)
High-quality coding assistant.
8-bit (MLX)
Smaller coding model; more iPhone-friendly and good for everyday coding help.