Launch in 5 Minutes
The Ideal AI Platform for SMEs
Supporting both cloud and on-premises deployment to perfectly balance secure governance, rapid implementation, and cost efficiency.
Empowering Businesses with True AI Competitiveness
AFS AI Hub is more than just a cloud API subscription; it is a platform that gives enterprises true ownership and control over their AI. Your data, models, and corporate knowledge run entirely within a fully secure and controlled environment.
Three Steps to Rapid AI Inference Deployment
No MLOps team required, no complex environment configurations—AFS AI Hub makes model inference as simple as installing software.
STEP 02
Select the ideal LLM from our model library and load it with a single click to start using it via API or the built-in UI. From installation to your first API call, go live with model inference in as little as 5 minutes.
STEP 03
As your business grows, seamlessly expand from the inference platform with a diverse range of applications from our AI Marketplace to continuously enhance your enterprise AI capabilities.
STEP 01
Simply sign up as a Taiwan AI Cloud member to download and deploy on-premises, or subscribe via the U-T Cloud. The single-package installation encapsulates the entire environment, eliminating tedious setups for an effortless start.
Six Core Advantages
Predictable AI Costs
Escape the unpredictable black hole of token-based pricing. Once implemented, your marginal costs continuously decrease as utilization increases, saving 60–80% compared to public clouds. Real-time monitoring of token consumption and GPU resource utilization gives administrators clear usage reports to easily track AI investment ROI.
High-Efficiency GPU Inference
Deeply optimized for cross-platform support across AMD, NVIDIA, and Qualcomm GPUs. Localized inference eliminates network latency while maximizing throughput and response speeds.
One-Click Model Access
Gain instant access to leading open-source LLMs including Gemma, GPT-OSS, Llama, and Nemotron. OpenAI-compatible APIs allow existing systems to integrate seamlessly, taking your inference live in no time.
Absolute Sovereign Control
On-premises deployment ensures that corporate secrets, customer data, and proprietary manufacturing know-how never leak to external platforms. Achieve zero risk of data exposure, perfectly meeting the stringent security demands of regulated sectors like manufacturing and healthcare
Modular Architecture with Agile Scaling
Start from a single inference model and gradually expand with diverse applications from the AI Marketplace as your business grows, ensuring you are never locked into a single vendor.
Localized Technical Support in Taiwan
Backed by Taiwan AI Cloud’s professional onboarding services. We possess deep expertise in Taiwan’s SME ecosystem and regulatory requirements, delivering rapid-response technical support in Traditional Chinese.
Flexible Deployment: Cloud or On-Premises
Whether you choose the cloud for rapid deployment or on-premises for ultimate control, AFS AI Hub delivers comprehensive support.
UT Cloud Solution
UT Cloud
Deploy AFS AI Hub via a UT Cloud subscription to enjoy a comprehensive cloud management interface and technical support services. Go live rapidly without the need to build your own infrastructure.
Full SLA Guarantee
Jointly backed by Taiwan AI Cloud and UT Cloud to deliver technical support, ensuring stable service operations and continuous version updates.
Flexible Subscription Pricing
Pay-per-hour billing paired with agile resource scaling, lowering the barrier to initial entry.
Zero setup, zero maintenance. Simply subscribe via UT Cloud to instantly access AFS AI Hub services—perfect for enterprises looking to rapidly pilot AI initiatives.
On-Premises Hardware Compatible Models
NVIDIA® GB10 Grace Blackwell
ASUS Ascent GX10
NVIDIA DGX Spark™
AMD Ryzen™ AI Max 300 Series (Strix Halo)
ASRock Industrial AI BOX-A395
HP Z2 Mini G1a
Qualcomm Cloud AI 100 Ultra
Aetina AIP-FR68-A2
Taiwan AI Cloudmembers can download AFS AI Hub for free to deploy on their own Edge AI devices, gaining absolute control over data and computing resources.
Product Overview Video
Watch this quick video to explore the core features and real-world application scenarios of AFS AI Hub.
AI Marketplace
Starting from the inference platform, gradually expand your proprietary AI applications. Our curated selection of applications is ready for immediate deployment, with more arriving soon.
AI Medication Safety Agent
Utilizes OCR and RAG technologies to index TFDA drug labels, automatically reviewing 28 prescription safety metrics to significantly boost pharmacist efficiency and medication safety.
RAG / OCR / Healthcare / Prescription Review
Applicable to: Pharmacies / Hospitals / Clinics
i-Spec Technical Document Agent
Upload technical specification documents to automatically extract terminology and build a context-aware RAG Q&A system, enhancing information retrieval reliability.
RAG / Technical Documents / Manufacturing / Knowledge Management
Applicable to: Manufacturing / Engineering / R&D
AFS MyAgent
Capable of executing commands, managing files, browsing the web, and integrating with LINE. It continuously expands its capabilities through skills and memory systems to automate various tasks for you.
AI Agent / Automation / LINE Integration / Task Execution
Applicable to: All Industries / Personal Productivity
Upcoming Categories
AI Enterprise Operations & Management
AI-assisted applications covering core operational workflows such as HR, finance, and procurement.
AI Agents
Proprietary AI agents for individuals and teams to boost daily productivity and knowledge access.
AI Vertical Integration Solutions
End-to-end AI solutions deeply integrated tailored for specific industries.
AI Content Creation
AI-powered generation and creation tools for multimedia content including copy, images, audio, and video.
AI Intelligent Customer Service
Multi-channel intelligent customer service systems to automate inquiries and elevate service quality.
AI Smart Manufacturing
Manufacturing AI applications integrating equipment monitoring, quality inspection, and predictive maintenance.
AI Information Security
Solutions for threat detection, anomaly behavior analysis, and automated security incident response.
AI Robotic Process Automation
Cross-system workflow automation that reduces manual tasks and minimizes error rates. AI City
AI City
AI-powered applications designed for smart city infrastructure, public services, and municipal governance.
AI Finance
Financial AI applications encompassing risk assessment, fraud detection, and intelligent robo-advisors.
FAQ
Who is AFS AI Hub built for?
AFS AI Hub is tailor-made for enterprises seeking rapid AI implementation while prioritizing data security and cost control. This includes:
• Manufacturing
• Healthcare & Pharmacies
• Finance & Insurance
• Retail & Customer Service
• SMEs & IT Teams
Even without a dedicated AI team, businesses can achieve rapid deployment and go live with applications effortlessly.
Is a GPU server strictly required?
Not necessarily.
AFS AI Hub supports a diverse range of deployment modes:
‧ Cloud Subscription (U-T Cloud)
‧ Edge AI Workstations
‧ On-Premises AI Servers
‧ Industrial AI BOX
Enterprises can select the optimal configuration based on model scale and specific use case scenarios.
How long does deployment take?
The basic inference environment can be up and running in as little as 5 minutes.
For standard enterprises, the entire transition from initial installation to actual application deployment can typically be completed within 1 working day.
Do I need a dedicated AI or MLOps team?
No.
AFS AI Hub comes pre-integrated with:
‧ Inference Engines
‧ Model Management
‧ GPU Environments
‧ API Services
General IT personnel can easily handle the entire deployment and management process.
Is the OpenAI API supported?
Yes.
AFS AI Hub provides an OpenAI-compatible API, allowing your existing systems and applications to integrate rapidly while slashing development and migration costs.
Which models are supported?
We currently support a wide variety of mainstream open-source models, including:
‧ Gemma
‧ Llama
‧ GPT-OSS
‧ Nemotron
Additional model versions can be continuously expanded and integrated over time.
Will my data be transmitted to an external cloud?
No. When choosing the on-premises deployment mode, all data handling, model inference, and platform operations are processed entirely within your enterprise infrastructure. No data is passed through external APIs or third-party clouds.
How do I get started with a trial?
You can:
‧ Request an AFS AI Hub Trial
‧ Contact a Consultant to Schedule a Deployment Evaluation
Recruiting AI Application Development Partners (ISVs)
We are actively inviting Independent Software Vendors (ISVs) to join us in building the AFS AI Hub application ecosystem.
Bring your AI applications to our Marketplace, reach Taiwan’s SME market, and co-create commercial value. Development tools, API documentation, and onboarding support are fully provided.