ARX-100 Edge AI Server
Deploy, use, and manage Private AI more simply.
The ARDGE ARX-100 is an integrated on-premises Generative AI appliance. Its compact, low-noise, energy-efficient design brings NVIDIA-powered AI to modern offices and edge environments. Powered by ARX OS, ARX-100 combines GPU computing, model deployment, and AI applications in one intuitive platform — helping organizations build Private AI faster while keeping models and enterprise data within their own environment.
ARX-100 features 128GB Unified Memory, with 96GB available for AI workloads, supporting a wide range of open-source models on a single compact system.
The graphical, no-code ARX OS interface simplifies model deployment and management, reducing setup complexity and the need for specialized AI infrastructure expertise. Supported models can be deployed locally in as little as 15 minutes.
ARX provides the tools to process enterprise data and run AI services locally — keeping sensitive data, intellectual property, and proprietary knowledge within your own environment.
Built for organizations that prioritize data sovereignty, privacy, compliance, and Air-Gap deployment.
Self-Service Knowledge Base & AI Agent Platform
A self-service knowledge base and AI chatbot for small and mid-sized businesses and non-technical teams.
Simply upload your documents. ARGENT prepares and organizes the knowledge, enabling teams to quickly build a natural-language chatbot without complex setup or AI expertise.
On-Premises Document Preprocessing
Process complex PDFs and images with mixed text, tables, and visual content, and convert them into structured Markdown.
High-quality data preparation is the foundation of accurate AI responses. Generate cleaner, AI-ready data for on-premises RAG, knowledge bases, and AI agents, with CLI and API access for external integration.
Turn Video and Audio into Searchable Knowledge
Extract speech from video and audio files, convert it into searchable transcripts, and import the content into AI agents.
Users can semantically search large media libraries, jump directly to relevant moments in a video, and transform accumulated video content into reusable organizational knowledge assets.
On-Premises AI Search for Large-Scale Documents
DocLake processes and organizes large volumes of documents, specifications, regulations, and forms into a private enterprise knowledge base. Users can search hundreds of thousands of documents in seconds using natural language — all on premises.
* DocLake is offered as a dedicated integrated appliance with ARDGE integration and FDE deployment services.
ARX App Market
Install, remove, and upgrade apps with a single click — simplifying deployment and maintenance. Extend from fully local knowledge AI to creative AI, digital workers, and more open-source applications.
A Hermes-based digital worker with an OpenClaw-like experience
ARX simplifies the installation, configuration, and security controls of open-source agents, making them easier to deploy and manage in a controlled enterprise environment.
Coming soon — December 2026 via ARX App Market upgrade.
Multiple generative AI workflows in one workspace
For strategic documents, long-form reports, complex analysis, and content generation that require stronger models or larger context windows, Hybrid Mode + De-Identification can securely extend workloads to Private Cloud, VPC, or Public Cloud AI. Keep sensitive knowledge local while gaining greater flexibility and compute capacity for creative AI workloads.
Coming soon — December 2026 via ARX App Market upgrade.
Private Templates + Private GitHub
Enterprise IT teams can deploy internally developed apps, agents, and vibe-coded workflows, then centrally publish and update them across multiple ARX devices through the App Market.
And more to come.
Supports popular open-source tools such as Open WebUI, AnythingLLM, Dify, and n8n, with more enterprise AI, agent, and automation applications continuously added.
ARX-100 can also serve as the on-premises AI foundation for third-party software and vertical solutions. With standard APIs, MCP, local model services, and containerized deployment, ISVs, SIs, and enterprise developers can focus on applications and industry use cases — without rebuilding the underlying AI infrastructure.
Preloaded and validated tools including Open WebUI, AnythingLLM, Dify, and n8n. Install via the App Market to build Private GPT environments and automated workflows.
ARX OS
A mobile App Store-like experience. Install and manage validated AI tools with one click—no command lines required.
Deep integration with the Hugging Face ecosystem. Supports visual downloads, inference management, and containerized deployment to ensure high environment consistency.
Provides standardized API endpoints for secure system integration. Built-in OTA enables one-click updates for the OS, apps, and models without on-site technical personnel, effectively lowering TCO.
A high-density AI workstation/server suitable for large model hosting and shared multi-department compute.
Supports external NAS or high-speed NFS architecture to accommodate growing document and media datasets.
Extend GPU resources beyond the edge to run larger models while maintaining absolute data sovereignty.
| Compute | |
| Architecture | NVIDIA® GB10 Grace Blackwell |
| Tensor Performance | Up to 1 PFLOP FP4 |
| System Memory | 128 GB LPDDR5x, coherent unified system memory (32 GB reserved for system, 96 GB available for AI workloads) |
| Storage | |
| Storage | 1 TB / 4 TB NVMe M.2 |
| Expansion | External NAS / high-speed NFS architecture |
| Connectivity | |
| Ethernet | 1× 10 GbE RJ-45 connector |
| NIC | ConnectX-7 NIC @ 200 Gbps |
| Display Connectors | 1× HDMI 2.1a |
| System | |
| Operating System | ARX OS |
| Physical | |
| Form Factor | Compact desktop edge appliance |
| System Dimensions | 150 × 150 × 51 mm (5.91 × 5.91 × 2.01 inch) |
| System Weight | 1.48 kg |
| Power Supply | 240 W |
| Acoustics | Ultra-quiet design |
| Compliance | |
| Certificates | BSMI / CB / CE / FCC / UL / CCC / C-Tick / WiFi / RF / VCCI |
Note: ARX-100 is a high-value, custom-configured AI system. Please contact ARDGE before purchase to confirm ARX OS features, specifications, and compatibility with your intended use. Custom-configured systems are non-returnable. Defective products remain covered under the applicable warranty and RMA policy.
| GUI-based Operation Console | |
| 1.1 | A browser-based graphical console centrally manages models, API endpoints, API keys, apps, GPU and system resources, logs, and storage mounts. |
| OpenAI & Anthropic Compatible API Endpoints | |
| 2.1 | Publishes models as standardized OpenAI- and Anthropic-compatible API endpoints, so development teams can integrate with existing SDKs and tools. |
| One-click Model Deployment | |
| 3.1 | Imports models with one click and automatically configures the inference engine and GPU resources. |
| AI Engine Platform | |
| 4.1 | Provides a model platform supporting multiple inference engines, on-premises deployment, and GPU resource allocation. |
| 4.2 | Centralized API key lifecycle management lets administrators issue, revoke, and expire keys by team or application, with RPM rate limiting and usage statistics. |
| Agent / MCP / CLI Integration | |
| 5.1 | Supports Agent, MCP, CLI, and API integration interfaces for connecting to existing systems, knowledge bases, databases, and workflows. |
| 5.2 | Provides AI Skills for built-in applications and integrates with mainstream agentic coding tools. |
| Model Usage Dashboard | |
| 7.1 | Provides a dashboard tracking per-model requests, token usage, response time, GPU utilization, and GPU memory usage across multiple time ranges (real-time / daily / 3-day / weekly). |
| Backup & Restore | |
| 11.1 | Backs up apps, models, and vector databases, with support for cross-machine (dissimilar-hardware) restore. |
Note: ARX-100 is a high-value, custom-configured AI system. Please contact ARDGE before purchase to confirm ARX OS features, specifications, and compatibility with your intended use. Custom-configured systems are non-returnable. Defective products remain covered under the applicable warranty and RMA policy.
| In the Box | |
| Main Unit | ARX-100 Edge AI Appliance × 1 |
| Power | Power adapter & power cord × 1 |
| Documentation | Quick start guide × 1 |
| Accessories | None |
Note: ARX-100 is a high-value, custom-configured AI system. Please contact ARDGE before purchase to confirm ARX OS features, specifications, and compatibility with your intended use. Custom-configured systems are non-returnable. Defective products remain covered under the applicable warranty and RMA policy.
Learn more about ARDGE’s latest Private AI solutions:
Download Flyer