Home
OFFICIAL RELEASE
Ex-Meta AI team -makers of the Hero Assistant app- launches SDK to supercharge natural language input, unlocking a wide range of use cases, a new form of advertising, and the foundation of a lightweight AR operating system.
SAN FRANCISCO – November 12, 2025. Today, the ex-Meta AI team behind the highly rated Hero Assistant app is announcing its patented AI Autocomplete SDK.
Fresh off an additional investment from Forerunner Ventures, CEO Brad Kowalk introduced the system, which has broad implications for how users interact with software. Hero will not only use this technology in its own assistant application, but will also allow other companies to use it for their applications via an SDK, unlocking up to 10X faster and cheaper actions.
AI Autocomplete is a real-time system that updates a user interface to guide people as they speak or type with everything needed to complete an action. This solves the core problem of natural language interfaces: people often don’t know what they need to say upfront to complete a task.
As an example, if someone wanted to book a flight with an assistant like ChatGPT, currently they would face a back-and-forth ‘game of 10 questions’, where the assistant would inquire about dates, times, origin, destination, luggage, seat preference, and more. The app may even open an in-app browser, leaving the user to watch it iterate in real-time.
“It’s slow, tedious, and the burden is on the user to think about what the AI needs,” says AI engineer Saharsh Vedi. “AI Autocomplete flips that model on its head.”
Instead, Hero's AI Autocomplete eliminates back-and-forth questions and browsing altogether, by front-loading all necessary information upfront.
“Hopefully we’ll look back and barely remember the concept of prompting," says Vedi.
Unlocks 10x Faster Actions
AI Autocomplete has the potential to transform any application that relies on natural language input across a wide range of use cases, for example:
Search & Commerce: Instead of manually searching and browsing, AI Autocomplete allows people to find -and ultimately purchase- exactly what they are looking for, including goods, tickets, food, and more. Here’s an example of ordering coffee online:
Media Generation: Instead of iterating through many versions of images or videos to find the perfect prompt, AI Autocomplete inspires users in real-time with parameters ranging from subject and action, to style, lighting, camera angle, and more. This not only increases the speed and creativity of users, but -just as important- significantly cuts server costs given far less iteration is needed.
Customer Service & Enterprise Forms: Today’s support chatbots use a series of questions and answers to get the right information from users to route and resolve tickets. Now, with AI Autocomplete, chatbots can guide users to provide all the information needed for their specific query, eliminating most back and forth questions. This not only saves time for users, but saves money for businesses that currently need to follow up to request missing information to resolve tickets.
Unlocks Natural Language Ads
In addition to transforming the user experience, AI Autocomplete also unlocks a new, intuitive way for businesses to monetize natural language interfaces.
First, similar to Google Ads, brands can pay to be ranked as a top suggestion during the decision making process. Second, brands can suggest complementary products to add to the current order. And finally, brands can benefit from higher repeat purchases, since recurring orders are as frictionless as a user saying “every month”.
For instance, as shown in the coffee ordering example above, Peet’s Coffee would not only benefit from being suggested as an option at the time of purchase, but also from recurring transactions.
Hero is currently in discussions with AI Advertising Platform Koah Labs to bring this new format to market.
Unlocks Powerful, Lightweight Augmented Reality
The inspiration for AI Autocomplete came from augmented reality. Co-founders Brad Kowalk and Seung Woo Lee first conceived of this technology after leaving Meta and discussing one of the big unsolved questions for AR adoption: how will AR devices be as useful as smartphones, if they have far less UI to guide users?
This directly contrasts with the heavy mobile-style UIs and gesture-based controls that large companies are investing in currently. Instead, AI Autocomplete would unlock a lightweight, powerful AR operating system driven by intent.
"We essentially had a realization that anything that you can do on the Internet is simply a series of parameters," says Kowalk. “It’s one of those things that -once you see it- is hard to unsee.”
SDK Brings Patented Technology to All Companies
To allow other companies to integrate this technology into their own products, Hero is announcing the AI Autocomplete SDK.
"We realized three years ago that the bottleneck to AI agents would not be AI’s ability to browse the internet and complete tasks, but rather the user's ability to know what to say," said CTO Seung Woo Lee. "So, instead of building another chat interface that plays a game of 10 questions, we are launching a system that lets users complete any multi-step action in one shot. We believe our technology has the ability to supercharge every product using natural language input today."
Companies interested in integrating the AI Autocomplete SDK into their software can reach out directly.