Run&Win is a Hong Kong-based AI technology company specializing in industrial robotics, smart health devices, and high-performance computing solutions. We serve clients across Asia-Pacific, Australia, and beyond.

Our product portfolio includes: E1 Underwater ROV for inspection, iRunYu Rail Track Bolt Oiling Robot for railway maintenance, H-Guard Smart Sleep Monitoring Pillow, and Pinyuan AI Compute Servers. Explore our full product lineup.

Yes. We work with enterprise clients to develop custom AI and robotics solutions tailored to specific industry needs. Get in touch with our team to discuss your project requirements.

The E1 Underwater ROV is rated for 100 meters (328 feet) maximum operating depth. This covers approximately 80% of infrastructure inspection scenarios including bridge substructures, dam faces, port facilities, inland waterways, and shallow coastal inspections. For deeper water applications, we offer custom configurations — contact our engineering team to discuss your specific depth requirements.

The E1 ROV has a 4-6 hour typical battery life depending on operating conditions (current, lighting, camera usage, speed). The system includes a quick-swap battery design so you can bring a spare and double your field time without returning to base. Battery level is displayed in real-time on the control console with low-battery alerts and automatic return-to-surface protection.

Each E1 ROV system includes:

• ROV vehicle with 4K UHD main camera + auxiliary camera
• 100m depth-rated tether
• Portable control console with built-in display
• 2 batteries + charger
• Transport case (wheelable, airline-checkable)
• Operation manual + quick start guide
• Basic training (4 hours online + 4 hours on-site or video call)
• 1-year warranty

Optional add-ons: laser scaler, multibeam sonar, CPT probe, extra batteries, extended warranty, on-site training. See full specs →

For most countries, no specific license is required to operate an inspection-class ROV like the E1 in inland and coastal waters. However, certain applications may require permits (e.g., working near critical infrastructure, protected marine areas). We provide a comprehensive operator training program and certification of completion with every purchase.

The E1’s 4K camera captures highly detailed imagery. With the optional laser scaler accessory, you can achieve ±1mm measurement accuracy for crack width, spalling dimensions, and deflection measurements. The laser scaler projects parallel laser lines a known distance apart, providing a precise scale reference in every image and video frame.

Yes. The E1 is designed for real-world infrastructure conditions. Its adjustable high-output LED array (up to 3000 lumens) combined with electronic image stabilization and low-light sensor performance allows operation in water with visibility as low as 0.5-1 meter. For extremely turbid conditions, we recommend adding the multibeam sonar option.

We offer comprehensive post-purchase support:

• 24/7 technical hotline for field emergencies
• 48-hour advance replacement warranty for critical components
• Free software updates for the data management system
• Annual maintenance service (optional service plan)
• Spare parts inventory shipped worldwide within 3 business days

Our support team includes certified ROV technicians and civil engineers who understand infrastructure inspection requirements.

The iRunYu Rail Bolt Oiling Robot operates at 3-5 km/hour along the track, oiling approximately 800-1200 bolts per hour depending on bolt spacing. For comparison, a skilled manual worker typically oils 200-300 bolts per day. That means one robot can replace 8-12 manual workers while delivering more consistent quality.

The iRunYu is compatible with standard gauge railway tracks (1435mm / 4ft 8½in) and most common bolt fastener types including:

• Spring clip fasteners (W-type, Nabla)
• Bolted rail joints (fishplate bolts)
• Screw spike systems

It works on both ballasted and slab track. For narrow gauge, broad gauge, or special fastener types, we offer custom configurations — contact us with your track specifications.

The iRunYu uses AI-powered computer vision with deep learning algorithms, achieving 95%+ bolt detection accuracy under normal conditions. The system automatically adapts to different bolt angles, lighting conditions, and track environments. If a bolt is missed (e.g., completely covered by debris), the system flags the location in the report so your team can follow up manually.

The iRunYu is powered by a lithium-ion battery pack providing 6-8 hours of continuous operation on a single charge. The battery is hot-swappable — swap in a fresh battery in under 2 minutes without tools. For long-distance jobs, bring 2-3 spare batteries and you can cover 30-50km of track per day with one robot.

The iRunYu is designed for low maintenance in demanding field conditions:

• Daily: Wipe down the vehicle, inspect hoses, check oil reservoir level
• Weekly: Clean the spray nozzles, lubricate wheel bearings
• Monthly: Full system diagnostic, calibration check, filter replacement
• Annually: Factory service (recommended) or on-site maintenance

All consumable parts (nozzles, filters, seals) are readily available and can be replaced by your maintenance team with basic training.

Absolutely. We offer on-site demonstrations at your rail facility, where we bring the iRunYu and put it through a real-world test on your track. You’ll see it oil bolts, generate a maintenance report, and can ask our engineer any questions. For qualifying customers, the demo is free of charge. Schedule a demo →

Yes — this is one of the biggest advantages over manual oiling. The iRunYu automatically generates a digital maintenance report for every job, including:

• Number of bolts oiled with GPS location
• Oil quantity applied per bolt
• Environmental conditions (temperature, humidity)
• Missed/damaged bolt locations flagged for follow-up
• Historical trend data for track sections

This data can be exported to your existing asset management system (CSV/SQL/API integration available) for predictive maintenance planning.

The biggest difference is zero wearability. H-Guard is built right into your pillow — you just sleep normally. No charging every few days, no uncomfortable band on your wrist, no ring that might get lost. The paper battery lasts 2-3 years without any intervention. While wearables measure through your skin (PPG optical sensors), H-Guard uses BCG (ballistocardiography) technology to detect the micro-vibrations of your heartbeat through the pillow. It’s completely passive and unobtrusive.

H-Guard’s heart rate monitoring has been validated against clinical ECG equipment with ±2 BPM accuracy, and HRV (Heart Rate Variability) measurements reach 98.2% correlation with reference ECG devices. Sleep staging accuracy is 85%+ compared to polysomnography (PSG), the gold standard for sleep testing. This is significantly more accurate than most consumer wearables and approaches clinical-grade performance.

Absolutely. We take data privacy seriously:

• All data is AES-128 encrypted both on-device and in transit
• You own your data — we never sell or share it with third parties
• You can delete all your data at any time with one click
• Family sharing permissions are granular — you control who sees what
• The system complies with GDPR, CCPA, and other major privacy regulations

For enterprise/B2B customers, we offer fully on-premises deployment options with no cloud connectivity required.

No — H-Guard works with or without Wi-Fi. Basic monitoring and data storage happen locally on the device. Wi-Fi is only needed for:

• Remote family viewing
• Cloud backup of historical data
• AI-powered health insights and reports
• Firmware updates

If you use it without Wi-Fi, data is stored locally and can be synced via Bluetooth when you open the app nearby.

H-Guard uses an ultra-thin paper battery (BD300300WAS, 2500mAh) with a rated lifespan of 2-3 years under normal use. That means no charging cables, no weekly battery top-ups — you literally set it up once and forget about it. When the battery eventually runs low, you’ll get a notification in the app, and replacement is straightforward (the battery module is user-replaceable).

Yes and no. The pillow itself can be shared, but each user needs their own profile in the app. The system can detect when a different person is using it based on weight and sleep patterns, but for accurate tracking, we recommend:

• Each person using their own pillow for consistent data
• Or, if sharing a bed, each person has their own H-Guard pillow

For couples, we offer a dual-pillow bundle with shared family account. The pillow is standard size (60×40cm) and fits most regular pillowcases. Learn more →

H-Guard includes 10+ health risk alerts that can notify family members or caregivers in real-time:

• Tachycardia (abnormally fast heart rate)
• Bradycardia (abnormally slow heart rate)
• Irregular heartbeat patterns
• Prolonged apnea (breathing pauses)
• Extended bed absence (may indicate a fall)
• Unusual sleep patterns

You customize who gets notified and for which conditions. Alerts can be sent via app push notification, SMS, or email.

PinYuan AI Compute Servers support 50+ mainstream AI models out of the box, including:

• LLMs: DeepSeek V3/R1 671B, Qwen 3.5 series, Llama series, GLM series, MiniMax
• Vision: YOLO v5/v8/v9/v11 series, PaddleOCR, various detection models
• Speech: FunASR, Whisper variants
• Multimodal: Qwen-VL, Z-Image-Turbo

New models are added regularly via the model marketplace. All models run through an OpenAI-compatible API — if your code works with OpenAI, it works with PinYuan with just one line of code changed.

The Jiangyuan D10/D20 AI accelerators in PinYuan servers deliver competitive performance, especially for inference workloads:

Metric	Jiangyuan D10	Jiangyuan D20	NVIDIA T4 (ref)
INT8 Performance	160 TOPS	320 TOPS	130 TOPS
FP16 Performance	80 TFLOPS	160 TFLOPS	16.2 TFLOPS
VRAM	64GB LPDDR5	128/256GB LPDDR5	16GB GDDR5
TDP	72W	150W	70W

Key advantages: much larger VRAM (single-card 256GB vs T4’s 16GB) and better performance per watt. For 70B+ parameter models that can’t fit on a single T4, a single D20 handles them comfortably.

Yes — PinYuan AI Compute Servers are built on a 100% domestic technology stack:

• AI Chip: Jiangyuan D10/D20 (fully domestic design + manufacturing)
• CPU Options: Hygon, Phytium, Kunpeng — all domestic
• Operating System: Kylin, UOS, openEuler
• Cloud Platform: Bingo CloudOS
• AI Models: Primarily Chinese models (Qwen, DeepSeek, GLM, etc.)

This makes PinYuan ideal for government, military, finance, healthcare, and other sectors where data sovereignty and domestic compliance are non-negotiable. Full certification documentation available on request.

In most cases, yes — with minimal changes. Here’s why:

• OpenAI-compatible API: Just change the endpoint URL and API key
• Standard frameworks: Supports PyTorch, ONNX, vLLM, and other standard tools
• Model marketplace: 50+ pre-optimized models, ready to run
• Migration support: Free migration assistance for enterprise customers

For custom models or specialized applications, we offer a 7-day proof-of-concept program to validate compatibility before you commit to purchasing.

PinYuan AI Compute Servers come in three product lines to cover all use cases:

Series	Form Factor	Use Case	Starting Config
Edge Series	Compact industrial PC	Edge AI, IoT, video analytics	1× D10, 160 TOPS
Station Series	Desktop workstation	Office AI, developer teams	1-4× D20, 320-1280 TOPS
Server Series	Rack-mount (2U/4U)	Data center, enterprise	2-16× Jiangyuan cards

All three tiers use the same software stack — develop on a Station and deploy on Servers without rewriting anything. Explore models →

PinYuan servers come with comprehensive enterprise-grade support:

• Standard warranty: 3 years hardware + 1 year software updates
• Extended warranty: Up to 5 years available
• Support channels: Phone, email, remote desktop, on-site service
• Response SLA: 4-hour response for critical issues, 24-hour for standard
• Training: Free administrator training + optional deep-dive technical training

Backed by Pingao Co., Ltd. (688227.SH), a publicly-listed cloud computing company with 20+ years of enterprise IT experience.

Yes, we offer several ways to evaluate:

• Cloud trial: Test PinYuan performance remotely through our cloud platform (free credits available)
• On-site demo: We bring a Station unit to your office for a live demonstration
• 7-day POC: Borrow a unit for 7 days to run your own workloads and benchmarks
• Enterprise pilot: 30-90 day pilot with up to 4 units for larger deployments

Contact our sales team to discuss which evaluation option makes sense for your organization.

Lead times vary by product and configuration. Standard products typically ship within 2-4 weeks. Custom solutions may require 6-12 weeks. Our team will provide a detailed timeline upon order confirmation.

Yes, we ship worldwide. International shipping rates and delivery times depend on the destination country and product dimensions. We’ll provide a shipping quote with your order confirmation.

All products come with a standard manufacturer warranty (1-3 years depending on product line). Extended warranty and maintenance packages are available for enterprise clients. Specific warranty terms are provided with each product quote.

Simply fill out the contact form with your requirements, and our sales team will get back to you within 2 business days with a detailed quote and product recommendations.

Yes, we offer tiered volume pricing for orders of 5+ units. Enterprise customers and resellers may qualify for additional discounts. Contact our sales team for a custom quote.