SINGAPORE – Singapore-based AI startup Estha today launched Estha on Mac, a private AI workforce for small and medium-sized enterprises and professionals. Built for Apple Silicon, it combines ready-to-use business agents with no-code tools for creating private AI applications, enabling supported confidential work to be processed on customer-controlled Macs rather than routinely sent to external AI inference services.
Estha on Mac is designed around recognisable business jobs, including meeting transcription and summarisation, confidential document analysis, presentation creation and spreadsheet analysis. This enables users without AI expertise to use AI tools out of the box. After installation and model download, supported local workflows can operate without an internet connection and without metered cloud-inference charges.
The launch comes as enterprises reassess how and where AI workloads should run. Broadcom’s Private Cloud Outlook 2026 found that 56% of enterprises were running or planning to run production AI inference on private cloud infrastructure, compared with 41% using public cloud for the same workloads. The report also found that 62% of IT leaders were very or extremely concerned about the infrastructure costs of generative and agentic AI.
“The work businesses most want AI to help with is often the work they are least comfortable uploading – the board paper, customer file, internal spreadsheet, contract or meeting recording,” said Eric Lam, co-founder and CEO of Estha. “Estha on Mac brings AI to that work on hardware the customer controls, through applications employees can use without becoming AI engineers.”
A private AI workforce for everyday business
Estha on Mac launches with four ready-to-use business agents:
• Meeting Intelligence transcribes and summarises meetings locally, extracting action items, decisions and key discussion points without sending recordings or transcripts to a cloud AI provider.
• Document Processing ingests and synthesises internal documents, contracts, pitch materials, research and other business content, helping users draft, refine and retrieve information in a controlled local environment.
• PPT Maker converts notes, documents and internal knowledge into presentation decks.
• Data Analyst analyses spreadsheets and internal datasets to identify trends, summaries and potential insights without exporting proprietary data to a third-party inference service.
Beyond the ready-made agents, Quick Bot and Advanced Studio allow users to create private chat apps from internal standard operating procedures, playbooks and knowledge documents without writing code. These customer-specific chat apps can support onboarding, training, internal knowledge and advisory workflows while running on the Mac.
Later this quarter, Estha will release two more agents: PDF Reader, which extends the workforce to confidential PDF analysis, and Vibe Coding: Prototype, which helps users turn an idea into an initial working prototype on the Mac.
Reducing cost, privacy and skills barriers to SME AI adoption
AI adoption among Singapore SMEs more than tripled in 2024, but remained at 14.5%, compared with 62.5% among non-SMEs, according to IMDA. The gap creates an opening for tools that are easier to deploy and use without specialist AI teams.
Broadcom’s 2026 survey found that cost had become the leading public-cloud concern among respondents, rising to 31% from 26% in the previous year. By eliminating metered cloud inference costs, Estha for Mac gives users the freedom to run AI as often as they need, without watching a usage meter.
As employees adopt AI faster than many organisations can approve and govern it, businesses face the risk of sensitive information being entered into unapproved services. Rather than relying only on restrictions, SMEs need sanctioned tools that are useful enough for employees to adopt.
Estha on Mac addresses one important part of this problem by processing supported data locally by default. This reduces routine external data transfers for those workflows and gives businesses greater control over where sensitive information is handled. Organisations remain responsible for appropriate access controls, device security, policies and human review.
Keeping sensitive work on the device
Cloud AI services can be configured with contractual, technical and organisational safeguards. But this requires businesses to understand how vendors handle data—including retention policies, access controls, model-training terms and cross-border data flows.
In light of the Personal Data Protection Commission’s guidance on the use of personal data in AI recommendation and decision systems, Estha for Mac’s on-device processing reduces exposure associated with transferring sensitive information to external AI services.
Estha for Mac is particularly relevant where privacy, predictable local costs and offline access materially affect adoption. Potential applications include:
• Healthcare: Clinics and health-tech companies can use local AI for administrative, documentation and knowledge workflows containing patient information.
• Financial services: Financial firms and financial consultants can apply on-device AI to internal documents, KYC materials, research and operational workflows with account, transaction and client information.
• Legal and professional services: Firms can process contracts, case materials, client documents and internal knowledge on-device.
• Education: Schools, training providers and education businesses can use local AI for internal administration, lesson preparation and knowledge workflows involving student or family information.
Mr Simon Lau, from United Christian College (HK), said: “We’ve used Estha on Mac for at least three use cases. It allows us to build AI workflows involving sensitive data like students’ grades, invigilation schedules without sending that information to an external AI service. The plug-and-play environment processes data locally — so tasks like analysing spreadsheets, documents and transcripts happen without confidential information ever leaving the device. Because users can also build and run their own AI applications locally, our teachers can use it without specialist AI knowledge.”
Mr Koay Tatt Chong, Director, Credence Advisory, said “At Credence Advisory, we have always believed that clients deserve bespoke advice, not cookie-cutter answers. The challenge is that as the firm grows, more time is inevitably spent on repeated internal queries, document preparation and operational workflows. Estha for Mac gives us a way to scale our internal expertise while keeping confidential information under control. More importantly, it allows our senior people to recover time for the work that matters most – applying expert judgement to serve clients better.”
Optimised for Apple Silicon
Estha on Mac is built for Apple Silicon and uses Apple’s MLX framework and Metal GPU acceleration. Apple Silicon’s unified-memory architecture allows the CPU and GPU to access a shared memory pool. For local AI applications, this can reduce the need to duplicate model data between separate system memory and graphics memory. Metal enables compute workloads such as matrix operations to be accelerated through the Mac’s GPU.
Availability
Estha on Mac can be downloaded from https://mac.estha.ai/
Media contact
For more information, contact
Betsy Tan, Director, Dopamine Communications
Email: betsy@betsytanpr.com
HP: +65 96416920
About Estha
Estha builds a private AI workforce for the Mac.
Built for Apple Silicon, Estha for Mac combines ready-to-use business Agents with no-code tools for creating private chat apps. Supported local workflows process documents and data on the Mac, reducing reliance on external AI inference services and allowing work to continue offline after installation and model download.
Its Agents support everyday work including meeting intelligence, document and PDF analysis, presentation creation, spreadsheet analysis and internal knowledge.
Estha’s mission is to make capable, private AI accessible to businesses and professionals by turning local AI into useful work on customer-controlled devices.
Go to https://mac.estha.ai/business for more information.
Via: technologynewsph
