See
See and understand
Read a sign, ask about a piece of equipment, or read contact details from a business card, with scene descriptions, text recognition and live visual coaching.
The app
A private AI assistant on your iPhone, through headphones or AirPods, from your wrist, or hands-free with glasses.
What it does
Ask about what you see, translate a conversation, remember the details, and get things done.
See
Read a sign, ask about a piece of equipment, or read contact details from a business card, with scene descriptions, text recognition and live visual coaching.
Talk
Ask a quick question or settle into a live voice conversation, translate spoken language and follow along with captions.
Remember
Save where you parked, keep notes about people and places, and return to saved conversations. The on-device brain connects useful context for later questions.
Act
Check your calendar, set a reminder, find directions, control your music or the lights, and connect Home Assistant, Siri Shortcuts and MCP tool servers.
Offline
With the on-device setup (local AI, speech recognition and voice), the conversation runs offline on your device.
Accessibility
Reading Assistant, Blind Assistant, spoken descriptions, captions, VoiceOver support and voice controls.
Make it yours
Give a persona its own wake word, model and prompt preset. Choose the AI, speech recogniser and spoken voice separately.
Capture
Capture photos and video, keep audio recordings for playback and transcription, or share a view with a configured remote expert connection.
Connect
Connect Home Assistant, Siri Shortcuts or compatible MCP servers, and add custom tools. Available actions depend on each service and its permissions.
What you need
The standard phone version is also usable on iPad and iPhone Duo. Dedicated versions for those devices are coming soon.
1
Running iOS 26 or later. That is enough on its own.
2
Spoken replies through headphones or AirPods connected to your iPhone or iPad.
3
The assistant from your wrist, using the iPhone paired with your Apple Watch.
4
Compatible Meta smart glasses connect through your iPhone to add hands-free camera and audio.
Your assistant, your choice
Pick the AI, speech recognition and spoken voice separately.
Use the provider you already have, with your own key.
Point Avenkin at a model you run yourself.
Apple Intelligence on supported devices, or compatible downloaded models running with MLX or llama.cpp. Choose on-device speech and voice for a conversation that needs no cloud API key.
Apple Intelligence uses Apple's on-device Foundation Models framework and needs no API key. It must be enabled and its system model ready. It supports text conversations and a selected set of enabled app tools, prioritising Field Assist actions when that mode is active. Camera understanding still needs a vision-capable model.
Avenkin is coming to the App Store. Until then you can join the beta through TestFlight, Apple's app for testing pre-release builds. The app's source is also published, so you can build it from source on a Mac.