Touchy, a voice and video-first AI assistant built by three UC Berkeley alumni, just launched from Y Combinator's Summer 2026 batch with a focus on making personal AI truly integrated into daily life. The app lets iPhone users activate an AI assistant with a single button press from their lock screen, then speak commands or show the camera what they're seeing. Touchy can then take action across more than 1,000 popular apps, access contacts and location data, and remember preferences over time to personalize itself to each user.

What Problem Does Touchy Actually Solve?

The founders identified a frustration many people experience with existing voice assistants. When you ask Siri or Alexa to send a message to a colleague, the assistant often misunderstands the request or returns search results instead of taking action. Users then have to manually open the app, find the right contact, and type the message themselves. Touchy was designed to bridge this gap by combining the intelligence of modern AI agents with the convenience of a device-native app that understands both what you're saying and what you're seeing.

The team believes that for AI to feel truly personal, it needs access to more than just digital tools. Your iPhone's camera and microphone capture rich information about the physical world around you, yet most AI systems today are limited to text and digital app interactions. Touchy taps into this multimodal capability, allowing the assistant to see a photo you snap and understand context from your environment.

Who Built Touchy and What's Their Background?

The three founders bring complementary expertise in AI research, education technology, and infrastructure engineering. Rohit Agarwal is a computer science PhD student at Princeton focused on AI alignment, and he graduated first in his class from UC Berkeley's EECS program in 2024. He previously worked at Duolingo and conducted theory research at Berkeley. Andrew Liu was a researcher at UC Berkeley working on language documentation tools and holds degrees in computer science, mathematics, and linguistics. Kevin Wang was a founding engineer at Eventual, a startup that raised $30 million in Series A funding, where he built data infrastructure for AI systems processing petabytes of data daily. He also worked at SkyPilot optimizing GPU training jobs and holds a degree in EECS from UC Berkeley.

The three met during their undergraduate years at Berkeley and are now reuniting to research and productionize what they see as the future of personal technology. Their combined background spans AI research, consumer product development, and the infrastructure that powers modern AI systems.

How to Get Started With Touchy

• Download the App: Touchy is available for free on the Apple App Store for iPhone users who want to try the voice and video-first assistant interface.
• Activate From Lock Screen: Press the Touchy button from your iPhone's lock screen to instantly start speaking commands without unlocking your phone or opening other apps.
• Build Your Preferences: Use Touchy in a continuous conversation where it automatically collects your memories and preferences, allowing it to personalize its responses and actions over time.
• Provide Feedback: Report bugs or share feature requests by shaking your phone or telling Touchy directly, or email the founders at founders@touchyapp.com to help shape the product.

What Sets Touchy Apart From Other AI Assistants?

Several design choices distinguish Touchy from competitors. First, it is built as a fully native iOS app with a custom interface designed to be fast and delightful to use, rather than a web-based or cross-platform tool. Second, it integrates with more than 1,000 popular apps out of the box, plus iOS-native capabilities like contacts and location services. Third, it maintains a single continuous conversation thread where it learns and remembers what you tell it, building a personalized model of your preferences and needs over time.

The founders emphasize that Touchy is designed to give users their attention back. Rather than fighting for screen time or notifications, the app is meant to be activated when you need it, complete your task, and then step out of the way so you can return to your day.

Why Does This Matter for the AI Startup Ecosystem?

Touchy represents a broader trend in Y Combinator's recent batches toward multimodal AI assistants that integrate deeply with consumer devices. The startup's focus on voice and vision, combined with app integrations and memory, reflects where the AI industry is moving: away from isolated chatbots and toward AI systems that understand context from multiple input types and can take real-world action. The founders' backgrounds in AI research, consumer products, and infrastructure also highlight how Y Combinator continues to attract talent from top academic and industry positions to build consumer-facing AI tools.

The app is now live and available for download, with the founders actively seeking user feedback to understand how people use Touchy and what features they wish existed. This early user engagement phase will likely inform the product roadmap as the startup moves forward from its Y Combinator launch.