Singapore-based Acti has secured US$5.3 million in seed funding to build what it calls an “agentic keyboard”, a product that aims to move AI assistance away from standalone chat apps and into the interface people use across almost every digital interaction.
The round was led by US-based BITKRAFT Ventures, an investor better known for backing gaming, interactive media, and consumer technology companies.
Acti said the capital will be used to hire engineering and AI talent, strengthen its on-device intelligence, and expand its ecosystem of Skills and developers.
Also Read: The coming identity crisis of agentic AI
The company’s core pitch is simple: the keyboard remains one of the few interfaces that cuts across messaging apps, email, productivity tools, browsers, and workplace software. Instead of asking users to open a separate AI application, paste in text, explain the context, and then move the output back to another app, Acti wants the AI layer to sit inside the keyboard itself.
That makes the product part of a larger race to define how AI agents will interact with users. While much of the current generative AI market has been built around chatbots and copilots, the next phase is expected to involve agents that can act across apps, remember preferences, and complete small recurring tasks with limited prompting.
A keyboard as the AI context layer
Acti’s product is built around programmable “Skill Keys”. A user can assign a function to any key, such as translating a message, generating a meeting link, rewriting a reply, summarising text, or triggering a workflow. These Skills can be created without coding. Users describe what they want through a Skill Builder, and Acti assembles the function.
According to the company, early access users created more than 1,000 Skills in under two weeks, suggesting demand for lightweight automation tools that do not require users to leave their current workflow.
Acti’s longer-term ambition is more ambitious than keyboard shortcuts. It wants to build a secure, user-owned, on-device personal context layer for the AI agent era. In practice, this means the keyboard would learn a user’s habits, preferred apps, frequently repeated tasks, and writing patterns over time, while keeping that knowledge on the device rather than inside a single application or platform.
Young Wang, CEO and founder of Acti, said today’s AI agents are limited because user context remains fragmented across apps. Acti’s cross-app presence, he added, gives it a chance to create a context layer that belongs to the user rather than any platform.
That positioning matters. The biggest AI companies are trying to pull users deeper into their own ecosystems. OpenAI has ChatGPT and its growing agent capabilities; Google is embedding Gemini into Android, Gmail, Docs, and Search; Microsoft is pushing Copilot across Windows and Office; Apple is integrating Apple Intelligence into iOS and macOS; and Grammarly is expanding from writing assistance into broader workplace AI.
Also Read: Agentic AI: The next frontier in technology
Acti is taking a different route. Rather than becoming another destination app, it is betting that the keyboard can become an AI distribution layer.
The Southeast Asian angle
Acti’s Singapore base gives the company a potentially useful launchpad. Southeast Asia is mobile-first, multilingual, and fragmented across consumer and business platforms, exactly the kind of environment where a cross-app AI interface could be tested at scale.
The region’s digital economy remains one of the world’s fastest-growing internet markets. Google, Temasek, and Bain & Company estimated Southeast Asia’s digital economy at US$263 billion in gross merchandise value in 2024. The region also has hundreds of millions of mobile internet users, many of whom move constantly between messaging apps, commerce platforms, ride-hailing apps, payment tools, and workplace software.
This creates a real pain point for AI products. A user in Singapore, Indonesia, Vietnam, or Thailand may communicate in multiple languages, switch between personal and work apps, and rely heavily on mobile-first workflows. Translation, rewriting, summarisation, scheduling, and message automation are not fringe use cases in this market; they are everyday productivity problems.
For startups, the opportunity is not only consumer adoption. Small businesses, sales teams, creators, recruiters, support agents, and cross-border sellers across Southeast Asia spend large parts of their day responding to messages, generating repetitive text, and coordinating across fragmented tools. If Acti can turn those behaviours into reusable keyboard-level Skills, it could find demand beyond early adopters.
The crowded AI productivity race
The challenge is that Acti is entering a crowded and fast-moving market. AI writing assistants, keyboard apps, and workflow automation tools are converging quickly.
Microsoft SwiftKey already integrates AI features. Google’s Gboard benefits from Android distribution and Google’s AI stack. Grammarly has strong brand recall in writing assistance and is moving deeper into enterprise productivity. Notion, Slack, Zoom, Canva, and Atlassian are embedding AI into their own workflows.
On the automation side, Zapier, Make, Raycast, and newer AI agent startups are also trying to reduce repetitive work.
The larger platforms have clear advantages: distribution, data, operating system access, and existing user accounts. Apple and Google, in particular, control the mobile operating systems on which third-party keyboards operate. This can limit how deeply an independent keyboard company can integrate, especially around privacy, permissions, and cross-app actions.
Acti’s answer appears to be user control and on-device context. If it can keep sensitive behavioural data on the device while still making the AI useful across apps, it may offer an alternative to platform-owned assistants.
This is, however, technically difficult. Personalisation requires data; privacy requires restraint; and agentic actions require reliability. A keyboard that makes mistakes is not just inconvenient; it can interrupt communication in the most visible part of a user’s workflow.
Also Read: The agentic economy: How to build a workforce where humans and AI collaborate
Jonathan Huang, Partner at BITKRAFT Ventures, said Acti reflects “an architectural shift” by reinventing the interface every app depends on and turning it into a layer AI agents will need.
From shortcut tool to agent infrastructure
The seed funding gives Acti room to prove whether its early engagement can translate into sustained usage. The near-term test will be whether users continue creating and using Skills after the novelty wears off. The longer-term test is whether developers see enough value to build around the platform.
If Acti succeeds, it could occupy an unusual position in the AI stack: not an app, not a chatbot, and not an operating system, but a persistent layer that follows users across applications. That would be particularly relevant in Southeast Asia, where users often operate across multiple languages, platforms, and commerce channels in a single day.
For now, Acti is still an early-stage startup with a bold interface bet. The US$5.3 million seed round shows investors are willing to back alternatives to the dominant AI assistant model. Whether the keyboard becomes the next control point for AI agents will depend on execution, privacy, and whether users are ready to trust their most frequently used interface with more than typing.
The post Singapore’s Acti raises US$5.3M to turn the keyboard into an AI agent layer appeared first on e27.