Touchy

Touchy is an iPhone-first personal AI assistant built by Tokenless, Inc. It combines voice and camera context with memory and actions across connected services. The current product and Y Combinator profile describe an assistant that can see what the user sees, hear spoken context, remember preferences and take actions across more than 1,000 app connections ¹². The App Store listing independently corroborates the iPhone product surface, iOS 18 requirement, camera-based assistance and broad connected-app positioning ³. These are unusually ambitious capabilities for a mobile assistant, but the quality and reliability of autonomous actions are not yet supported by broad independent testing.

The app is free to download with in-app purchases, and the terms state that new users receive a seven-day trial before a monthly or annual subscription renews ⁴³. A stable globally applicable public subscription amount was not established from the current first-party terms, so cost comparisons should be treated cautiously. Access is also materially constrained by the Apple ecosystem: the main product is designed for iPhone and requires iOS 18 or later ³.

Touchy has clear current first-party legal and privacy documentation. Its privacy notice says the service may share inputs, outputs and personal information with AI service providers including OpenAI, Google Cloud AI and Anthropic, and describes retention, account-deletion and security practices ⁵. Apple’s developer-provided privacy labels indicate that categories including location, contact information, user content, identifiers, usage data and diagnostics may be linked to the user ³. The privacy surface is therefore materially more consequential than for a text-only assistant because Touchy can receive camera, audio and connected-service context. The exact App Store listing currently lacks a sufficient ratings sample for an aggregate, so there is no eligible independent numeric user-rating signal ³.

Touchy is best suited to iPhone users who value hands-free, context-aware assistance and are comfortable granting an assistant access to camera, audio and connected services. Its native interaction model is especially relevant for quick real-world questions, inbox/calendar actions and other tasks where opening individual apps creates friction ¹². It is a weaker fit for users who need Android or desktop-first access, a mature enterprise-control surface, or extensive independent reliability evidence.