AskAIs AI: Surpassing One Million Total Users in 2025

The first phase of generative AI was about astonishment: AI could converse, write, draw, even generate code. In the next phase, the question users really care about has changed: can AI understand the context of work, connect different tasks, and help people actually get things done?

AskAIs AI grew up in the middle of this shift.

Launched in Hong Kong on 23 February 2023, it entered the market with GPT-powered conversation and Stable Diffusion image generation. Three years on, AskAIs has grown from an AI chat product into an all-in-one work platform integrating receipts, documents, task management, AI agents, image creation, team collaboration and a developer API.

Behind AskAIs, Stellar AGI Labs is also building a complete technology stack spanning AI applications, large models, APIs, enterprise software, and even compute and storage hardware.

This article is compiled from the official websites of Stellar AGI Labs and AskAIs as publicly available up to June 2026. User numbers, API call volumes and technical capabilities are the company's publicly stated figures.

A Startup Story That Began in Hong Kong

Stellar AGI Labs was founded by seven co-founders. The team began building during their university years; its members are spread across Hong Kong, mainland China and the United States, responsible respectively for product, engineering, research, operations and regional compliance.

"Stellar" represents the company's Hong Kong headquarters, "AGI" represents a long-term vision for artificial intelligence, and "Labs" represents continuous R&D and product experimentation. The company's core purpose is to raise human productivity through AI, and to help Hong Kong become a hub of intelligent technology.

Ambitious as that vision is, the team chose a very pragmatic starting point: first turn rapidly advancing large-model capabilities into a product that ordinary users can use directly.

2023: Launching the First AI Product

In February 2023, Hong Kong Stellar Digital Technology Limited was formally established. That same month, the team launched its first AI product, chat.stellar.hk, offering AI conversation and text-to-image generation.

At the time, the market was still debating whether AI could answer questions, and the first generation of AskAIs likewise focused on lowering the barrier to using large models.

But the team did not stay a chat product for long. In June 2023, Stellar released its first fine-tuned large model and opened an API, beginning to help enterprises and developers bring AI into their own products.

This step gave AskAIs two product lines: one for individuals and teams, solving "how to use AI"; the other for enterprises and developers, solving "how to make a product AI-capable".

2024: Entering the Mainland China Market

In 2024, the team established Shenzhen Xingwen Artificial Intelligence Co., Ltd. in Shenzhen, responsible for mainland China operations, algorithm filing, content safety and related compliance.

AskAIs also built a separate edition for mainland users, aligned with local models, payments, network environment and regulatory requirements. The official website currently displays large-model algorithm filing, ICP filing and value-added telecom business licence information.

This shows that AskAIs' internationalization is not merely translating an interface, but building region-specific product, operations and compliance capabilities.

2025: From Software to Compute and Storage

In 2025, Stellar began extending its technology map into AI compute and storage hardware. The company established a chip-technology company in Shenzhen for manufacturing, and set up an R&D headquarters in Hong Kong.

Its first hardware product, AskAIs Mini SSD, uses the M.2 NVMe 2280 form factor, offers 128GB to 4TB of capacity, and is positioned for large-model training and related high-throughput storage scenarios. The official product page indicates the product is research-complete and planned for launch in Hong Kong.

This means Stellar hopes to form three layers of AI capability: AskAIs AI serving users directly; AskAIs LLM and the API delivering intelligence to enterprises; and compute and storage hardware extending into the underlying infrastructure.

That said, hardware and SaaS are two entirely different industries. An SSD must ultimately prove its value through its controller, NAND, firmware, endurance, thermals, sustained read/write, mass production and after-sales capability.

2026: Building a Three-Region Operating Structure

In 2026, the team established MiniCode LLC in the United States, responsible for international-market operations and global compliance design.

With this, AskAIs forms a three-region operating structure spanning Hong Kong, Shenzhen and the United States: Hong Kong is the group headquarters and the gateway to Asia; Shenzhen handles mainland China product and compliance; and the US company takes on the international market.

This structure helps handle different regions' data, payment and regulatory requirements, but it also means the platform must more clearly explain the service entity, where data is stored, and cross-border transfer arrangements.

From a Chat Tool into an AI Work Platform

Today's AskAIs is no longer just a chat interface; it integrates multiple stages of work into one platform.

AI receipt and invoice features can generate documents from natural language, extending to batch generation, automatic sending, multi-currency, multi-language and API access. This is a concrete, high-frequency entry point, because a receipt is not just text — it involves fields, amounts, counterparties and delivery formats.

AI smart documents handle content generation, classification, retrieval, storage and sync. They solve not only "how to generate a document" but also how to find, update and share it afterwards.

AI Todo can break a larger goal into tasks, set reminders and track progress, shortening the distance from discussion to execution.

AI Agent attempts to move from passive Q&A to continuous assistance, supporting multi-turn dialogue, search, creation and reasoning. Agent is still in Beta; its long-term value will depend on whether it can safely call tools, advance tasks, and hand decisions back to the user when uncertain.

The platform also offers image and poster generation, plus a team shared space, document sync and permission control. These capabilities let AskAIs evolve from a personal productivity tool into a team work system.

A Complete Workflow

Suppose a Hong Kong consulting firm receives an inquiry from an overseas client. An AI Agent can first organize the requirements, then generate a service proposal and collaboration documents.

The Todo system can split the project into research, production, review, delivery and follow-up tasks; the image feature can create proposal visuals; and team members collaborate in the shared space.

Once the project is complete, the platform can generate a receipt or invoice. If the enterprise already has a CRM, ERP or SaaS system, it can also plug these capabilities into existing workflows via the LLM API or Invoice API.

The real value of an all-in-one platform is not putting many features on one website, but letting client data, project context, documents, tasks and permissions flow across the entire work cycle.

A Three-Layer Product Architecture and Four Business Models

From public information, AskAIs is forming a three-layer architecture.

The first layer is applications for individuals and teams, including documents, receipts, Todo, Agent, images and collaboration; the second is the LLM API, Invoice API, model fine-tuning and customization for enterprises and developers; the third is the model, compute and storage infrastructure now being built.

Its business models likewise span a free entry point, individual and team subscriptions, pay-as-you-go API billing, and enterprise customization and private deployment.

This model can build a user base through free products, then raise customer value with subscriptions, API and enterprise services. But different customers have entirely different requirements: individuals value simplicity, developers value documentation and stability, and enterprises care about security, permissions, service levels and contractual guarantees.

A Million Users and Ten Billion API Calls

According to figures published by Stellar AGI Labs for 2025, AskAIs surpassed 1 million total annual users, and cumulative API calls exceeded 10 billion.

User numbers reflect the front-end product's market reach, while API call volume reflects high-frequency machine-to-machine use. Ten billion API calls do not equal ten billion human interactions, but they show that AskAIs' usage does not come from the chat interface alone.

The metrics worth watching in the next phase will be monthly active users, paid conversion, team retention, the number of enterprise API customers, service latency and task completion rate. What the market ultimately cares about is not how many people have used it, but how many keep getting work done through the platform.

Security and Data Governance

AskAIs' privacy policy states that primary data is stored in the United States, and that some data of Asia-Pacific users may be processed via the Hong Kong operating region. After account deletion, personal data is in principle deleted or anonymized within 30 business days.

The platform also states that if data may be used to optimize model training, separate explicit consent will be obtained, and only de-identified data will be used. Its publicly stated security measures include TLS 1.3, AES-256, tiered permissions, data masking, intrusion detection and third-party security audits.

For formal enterprise adoption, it is still necessary to further confirm the scope of security certifications, data processing agreements, sub-processors, data residency, audit logs and incident-notification mechanisms. The deeper AI goes into core business, the more transparency matters.

Opportunities and Challenges of a Full-Stack Path

Stellar AGI Labs' business spans AI, IT and Web3. This breadth has the potential to create synergy: enterprise software supports AI adoption, the API brings capabilities into other products, and hardware attempts to solve underlying supply issues.

But models, SaaS, chips, SSDs and a trading platform are each enough to be a standalone company. For a growing team, the biggest challenge is keeping priorities clear — clearly distinguishing products that already create user value, long-term R&D projects, and parts that require partners to complete.

AskAIs should remain the core of the entire product map. It is closest to users and best able to validate real needs. Defining model, API and hardware capabilities backward from AskAIs' work scenarios is more likely to form a sustainable product loop than simply expanding the technology map.

Conclusion: A Real AI Product Gets Things Done

AskAIs' three-year history is a path from application to platform, from Hong Kong to the world.

It started in 2023 with conversation and image generation, building models and an API that same year; in 2024 it built mainland China operations and compliance; in 2025 it extended into compute and storage; and in 2026 its US company took on the international market.

The core of this path should not be merely having more AI features, but turning large-model capabilities into workflows that individuals, teams and enterprises can keep using.

The future of AI is not just a chat box that answers questions better. It should understand goals, connect tools, respect permissions, retain context, and get things done under human supervision.

That is the direction most worth holding to as AskAIs goes from Hong Kong to the world.