Acti is betting that the next everyday home for AI agents is not another chatbot app. It is the smartphone keyboard, the interface people already use across email, messaging, social media, and many other mobile apps.
On Tuesday, the Singapore-based startup launched an agentic keyboard for iOS and Android. The product is designed to do more than suggest the next word: it can help users take actions without leaving the app they are already using.
Why Acti starts with the keyboard
The core idea behind Acti is simple: people often need AI help while they are already inside another app. A conversation may require a recommendation, a translation, a meeting link, or a quick piece of information. Today, that usually means leaving the conversation, opening another tool, getting the answer, and then returning.
Acti, short for “action,” tries to put that assistance directly where typing happens. If a friend asks where to eat nearby, the keyboard could insert a local recommendation. If someone mentions a stock in a chat, Acti could help share the live price in that same conversation.
Young Wang, Acti founder and CEO, frames the keyboard as a way to reduce the fragmentation that comes from jumping between separate apps. In an email interview with TechCrunch, Wang said, “Today’s AI agents are fundamentally limited because user context stays fragmented across separate apps.”
He said Acti “sits across all of them, which is why we can build a context layer that genuinely belongs to the user instead of the platform.” For Acti, that cross-app position is the strategic reason to focus on the keyboard rather than building another standalone AI destination.
How Skills turn typing into action
Acti is powered by Google’s Gemini models. Wang said Gemini was chosen for its balance of intelligence, speed, reliability, multilingual performance, and cost-efficiency.
One of the main features enabled by that foundation is called Skills. Skills work like custom keyboard shortcuts that can trigger multistep tasks. Instead of manually repeating a workflow, a user can assign an action to a single key.
The app ships with built-in Skills. One example is “T,” which lets a user translate a message to another language by long-pressing the letter on the keyboard. Another is “C,” which can send a meeting link.
Users do not need to code to create Skills. Acti says a user can describe the desired action in plain language, and the app builds the Skill. Ahead of launch, early access testers built over 1,000 Skills in less than two weeks.
Skills can remain private, or they can be shared publicly in a Skills marketplace. The marketplace can include Skills other users have created, including ones for accessing real-time World Cup data or Polymarket links. Acti also sees the future Skill Hub as a possible source of additional monetization opportunities.
Privacy is built around a local-first model
Because a keyboard sits inside sensitive daily communication, Acti’s privacy model is central to the product. The company says it is built around a local-first approach, meaning users’ personal context stays on their device by default.
Acti says the app does not access or store private messages, conversations, or personal context unless the user explicitly invokes a feature that requires external processing. That distinction matters because the keyboard’s value depends on context, while user trust depends on clear limits around when that context leaves the device.
The company’s positioning also reflects a broader question for consumer AI: whether people will adopt AI by opening dedicated bots, or whether AI will work best when embedded into familiar interfaces. Acti is built around the second view. It places the agent inside the flow of typing, where intent is already being expressed.
The founder’s bet on the AI-era keyboard
Wang’s focus on keyboards follows earlier experience at Baidu, where he spent a decade growing Facemoji Keyboard to over 300 million daily active users. He said the arrival of LLMs changed his view of what typed text could become.
“When LLMs arrived, I realized something fundamental had changed,” Wang said. “Text was no longer just something people typed; it had become a carrier of intent. And in many everyday contexts, that intent can now be directly translated into action.”
That view explains why Acti is treating the keyboard as more than an input field. If typed language can represent what a user wants to do, then the keyboard can become a place where that intent is translated into a task.
Wang added that it was time to reinvent “one of the most basic and universal products people use every day: the keyboard.” For Acti, the familiar keyboard becomes the surface where AI agents can move from answering questions to completing small actions.
Funding, team, and business model
Acti’s business model is still taking shape. The company plans to generate revenue through subscriptions that offer more advanced AI models, higher daily usage limits, and other premium features.
The startup also told TechCrunch exclusively that it has closed $5.3 million in seed funding. The round was led by BITKRAFT Ventures.
Jonathan Huang, partner at BITKRAFT Ventures, said the firm backed Acti because the team has “a real shot at owning the next phase of human-computer interaction.”
The company’s leadership includes CTO Mike Sun, who was the founding technical lead behind Yike Album, Baidu’s cloud-photo platform, which scaled to over 10 million daily active users. Acti’s CSO is Junbo Yang, who joined from HashKey Capital, where Yang led dozens of consumer investments.
The launch puts Acti into a practical test of whether users want AI agents inside the keyboard they already rely on. If the product works as promised, its value is not just faster typing. It is reducing the distance between writing something, understanding the context, and taking the next action.