f in x
Acti Launches AI Smartphone Keyboard That Performs Actions on Behalf of the User
> cd .. / HUB_EDITORIALE
News

Acti Launches AI Smartphone Keyboard That Performs Actions on Behalf of the User

[2026-07-01] Author: Ing. Pietro Maiorana
Zenithby Meteora Web The operating system for your business. Social, clients, bookings and invoices in one platform. Gyms, barbers, professionals. Discover Zenith Free demo · no card

A Singapore-based startup called Acti has introduced an intelligent keyboard for iOS and Android that goes beyond text prediction. The new agentic keyboard is designed to integrate with everyday applications such as email, messaging, and social media, executing actions automatically upon user request. The stated goal is to eliminate the need to switch between apps to get AI assistance.

The keyboard becomes a contextual AI assistant

According to Young Wang, founder and CEO of Acti, the main limitation of current AI agents is that user context remains fragmented across separate applications. By positioning itself as a cross-app layer, Acti's keyboard builds a context that belongs to the user rather than the platform. Wang, who previously spent a decade at Baidu growing the Facemoji keyboard to over 300 million daily active users, stated that the advent of large language models has transformed text from simple input into a carrier of intent, making it possible to directly translate intentions into actions.

Sponsored Protocol

Customizable Skills without coding

One of Acti's core features is Skills, customizable shortcuts that users can create by describing what they want in natural language. For instance, long-pressing the 'T' key instantly translates a message, while the 'C' key generates a meeting link. The company reports that during early testing, users created over 1,000 Skills in less than two weeks. These shortcuts can be private or shared publicly in a marketplace called Skill Hub, which may offer future monetization opportunities for creators as well.

Local-first privacy and Gemini models

Acti is built on a local-first model: personal data stays on the device by default, and the app does not access or store private conversations unless the user explicitly invokes a feature requiring external processing. Under the hood, the keyboard uses Google's Gemini models, chosen according to Wang for their balance of intelligence, speed, reliability, and multilingual performance. This positioning allows Acti to deliver fast and relevant responses without compromising privacy.

Sponsored Protocol

$5.3 million seed funding round

Exclusively disclosed to TechCrunch, the startup has closed a $5.3 million seed round led by BITKRAFT Ventures. Jonathan Huang, a partner at the fund, stated that Acti's team has a real shot at owning the next phase of human-computer interaction. The team includes CTO Mike Sun, former founding technical lead of Baidu's Yike Album platform, and CSO Junbo Yang, who joined from HashKey Capital. The business model involves subscriptions for advanced features such as more powerful AI models and higher daily usage limits.

Sponsored Protocol

For those interested in creating custom AI assistants without coding, Acti joins a landscape already rich with initiatives like ChatGPT Custom GPT, which enables building tailored assistants. Moreover, deep AI integration into mobile devices echoes strategies of companies like Apple and Google, as highlighted in the article on Zurich as a secret R&D hub. For further reading on virtual keyboards, see the Wikipedia page on virtual keyboards.

Source: https://techcrunch.com/2026/06/30/acti-puts-ai-agents-directly-into-your-smartphone-keyboard

Ing. Pietro Maiorana

> AUTHOR_EXTRACTED

Ing. Pietro Maiorana

Ingegnere informatico e co-fondatore di Meteora Web, CMO dell'agenzia. Esperto di marketing digitale, social media, advertising, copywriting e SEO.
[ Read Full Dossier ]

> METEORA_WEB // DIGITAL AGENCY

We build the digital presence your business deserves.

Websites, social media, online advertising, e-commerce and high-performance hosting, engineered with method by computer engineers in Sciacca, for all of Italy.

> MW_JOURNAL

> READ_ALL()