Home

Monako Glass AI Coding Glasses Review

Introduction: When Glasses Become a Linux Computer

Monako Glass AI coding glasses side view

On June 3, an obscure Chinese startup named Monako dropped a bombshell: Monako Glass, the world’s first wearable Linux computer in glasses form. Not an AI assistant hanging on your glasses—the glasses themselves are a complete computer. MonoOS system, Lua application layer, Claude Code and OpenAI Codex running locally. Forty-eight grams, $399, $19 reservation.

Founder Candy Yue said something striking in the launch video: “I used to grind at the keyboard, but nowadays I just tell the computer what I want, and the AI builds it for me.” This sentence precisely captures Monako Glass’s ambition—not letting you operate a computer with glasses, but making the glasses themselves your AI coding workstation.

Product Overview: A Niche Geek’s Crazy Experiment

Monako is a Chinese startup with unclear team background, but its product vision is crystal clear: jump out of the homogeneous AI glasses track of “camera + music + voice assistant” and directly target developers and researchers. Candy Yue’s demo video on X shows users generating applications, executing research tasks, creating presentations, and even converting handwritten formulas to LaTeX code in real-time through voice commands.

This positioning is extremely niche yet extremely precise. In the era when AI agents evolve from “code completion” to “autonomous debugging, code review, and automated workflows,” developers do need a more lightweight way to supervise AI completing tasks. Monako Glass wants to move this “supervision desk” from the desktop to your nose bridge.

Specifications: How Much Crazy Fits in 48 Grams

Spec	Details
Operating System	MonoOS (custom Linux)
Application Layer	Lua (200-500KB memory footprint)
Weight	48g
Display	Waveguide display
Camera	HD camera
Speaker	Built-in speaker
Microphone	Bone-conduction (nasal vibration capture)
Gesture Interaction	Vision Engine (0.5 TOPS NPU)
AI Support	Claude Code, OpenAI Codex
Interaction	Voice commands + gestures + heads-up display
Price	$399 ($19 reservation)
Estimated Shipping	July-August 2026

Data source: Monako official launch materials, Digital Trends, Times of AI

Monako Glass AI smart glasses front view

MonoOS is Monako’s self-developed Linux distribution, with a core feature of being lightweight—the Lua application layer uses only 200-500KB of memory. This means applications can still run smoothly on resource-constrained embedded devices. In comparison, a typical Android app consumes dozens of MB of memory, making MonoOS’s lightweight design enable running complex AI agents on glasses.

The bone-conduction microphone design is clever: placed near the nose rather than the ear, capturing voice through vibrations of bones near the nasal cavity. This design effectively separates user voice from background noise in noisy environments like coffee shops and offices, more precise than traditional microphone arrays.

The Vision Engine gesture system relies on a 0.5 TOPS NPU, supporting operations like raising hand to summon menus, tapping to select apps, and sliding to adjust volume. Half a TOPS is not powerful, but sufficient for simple gesture recognition, and the power consumption is low enough—crucial for battery-sensitive glasses devices.

Deep Analysis: Can It Actually Write Code?

Monako Glass’s core selling point is “AI agent running 24/7.” The official demo shows several enticing scenarios:

Voice-Driven Development: Say “create a to-do app,” and the AI agent generates code, debugs, and builds in the background, notifying you via vibration or voice prompt when done. You can check progress on the heads-up display and intervene via voice when necessary. This workflow eliminates the need to sit in front of a computer screen, suitable for “asynchronous programming” during commutes, waiting in lines, or even walking.

Handwritten Formula to LaTeX: Researchers write formulas with pen and paper; the glasses camera captures and converts them to LaTeX code in real-time. For math, physics, and engineering researchers, this saves significant manual input time.

Cross-Device Collaboration: Monako Glass is not an isolated device. It can serve as a “front end” for Mac or PC, with agents running on cloud or local servers while the glasses only handle input, output, and status display. This “light front-end + heavy back-end” architecture solves the glasses’ computing power limitation.

Monako Glass waveguide display Linux interface

But problems are equally obvious:

Battery Life Unannounced: The 48-gram weight suggests an extremely limited battery. If AI agents run 24/7, battery life might be only 1-2 hours. The fact that official battery specs remain unannounced is a red flag.

Chip and Storage Unknown: Processor model, memory size, and storage capacity are all unannounced. Running Linux plus AI agents requires at least 4GB memory and a mid-range Arm chip—whether the 48-gram body can fit these hardware components is questionable.

Display Quality Uncertain: A waveguide display in a 48-gram body likely has limited resolution and brightness. Will reading code for long periods cause eye fatigue? Is it readable in bright outdoor light? All unknown.

Privacy Paradox: Monako emphasizes “privacy first” because AI agents run locally without data leaving the device. But the camera on the glasses always exists. Wearing a Linux computer with a camera in public spaces raises social acceptance issues. Monako has not yet clarified whether the camera has physical遮挡 or LED indicators.

Comparison: Monako Glass vs Meta Orion vs Apple Vision Pro

Feature	Monako Glass	Meta Orion AR	Apple Vision Pro
Weight	48g	100g (prototype)	600-650g
Price	$399	Unannounced (premium expected)	$3,499
Positioning	AI coding workstation	Social + AR display	Spatial computing platform
OS	MonoOS (Linux)	Custom Android	visionOS
AI Capability	Claude Code/Codex local	Cloud AI	Cloud + local hybrid
Display	Waveguide	70°FOV transparent	4K Micro-OLED
Camera	Yes (privacy policy unclear)	Yes	Yes
Target Users	Developers, researchers	General consumers	Professional users + early adopters

Monako Glass takes a completely different path. Meta and Apple are building “next-generation general computing platforms,” while Monako is building “specialized tools for specific scenarios.” At $399—one-ninth the price of Vision Pro and one-thirteenth the weight—but functionality is also much narrower. It doesn’t do AR games, spatial video, just one thing: letting developers supervise AI agents writing code anytime, anywhere.

This strategy is smart. In the AI agent era, “human-computer interaction” is shifting from “humans operating software” to “humans supervising AI completing tasks.” Monako Glass bets that: in the future, not everyone needs a high-performance computer, but every developer needs a lightweight “agent supervision terminal.”

Pros and Cons

Pros	Cons
World’s first glasses-form Linux computer, concept ahead of its time	Core specs like battery, chip, storage unannounced
48g relatively lightweight for extended wear	Waveguide display quality uncertain, code reading experience unknown
Local AI agent execution, good privacy protection	Camera privacy issues not clearly addressed
Bone-conduction microphone, clear voice in noisy environments	$399 pricing relatively high for a startup brand
Voice + gesture interaction, hands-free	Extremely weak ecosystem, nearly zero third-party apps
Precise targeting of developer pain points	Shipping July-August, actual experience to be verified

Who Should Buy

Recommended for:

Developers who frequently need “asynchronous programming” (advancing projects during commutes or waiting)
Researchers needing to capture inspiration and convert formulas anytime
Early AI agent adopters willing to try new interaction forms
Privacy-sensitive users who don’t want code data on the cloud

Not recommended for:

Professional developers needing long immersive coding sessions (screen too small)
Users demanding brand stability and after-sales service (startup risk)
Budget-constrained users—$399 can buy a decent laptop
Socially sensitive users unwilling to wear camera-equipped glasses

FAQ

Q: Can Monako Glass replace my laptop?

A: No. It is better suited as an “agent supervision terminal” rather than a “primary workstation.” Complex debugging, large-screen code review, and multi-window operations still require returning to a desktop environment.

Q: Can MonoOS install regular Linux software?

A: Official compatibility is unclear. MonoOS is based on the Linux kernel, but the Lua application layer means regular Linux applications need porting or redevelopment.

Q: Is the $19 reservation a deposit or full payment?

A: The $19 is a reservation fee, not full payment. Specific final payment timing and refund policies need to be confirmed with official channels.

Q: Does it support prescription lenses?

A: Official information has not been announced. The 48-gram body has limited space, and custom lenses may add weight and thickness.

Conclusion

Monako Glass is a device where “concept exceeds product.” It raises a highly forward-looking question: when AI agents can autonomously complete most coding tasks, what hardware do developers need to “supervise” this process? The answer may not be bigger screens or stronger CPUs, but a lightweight pair of glasses that keeps you connected to your agent in any scenario.

But a forward-looking concept doesn’t mean a mature product. Battery, display, ecosystem, privacy—these fundamental questions remain unanswered by Monako. At $399, the pricing carries no small risk for a first-generation product from a startup.

If you are a developer willing to pay for future interaction forms, Monako Glass is worth a try. But if you need a reliable “works right now” productivity tool, I recommend waiting for first-batch user feedback in July-August before deciding.

2026-06-06

Steam Frame VR Headset Review

Introduction: Valve Finally Returns

In 2019, Valve Index burst onto the scene with its 144Hz refresh rate and Index controllers’ “finger tracking,” becoming the ceiling of PC VR at the time. Then Valve disappeared—no new VR hardware for six years, watching Meta Quest evolve from a 3DoF toy to the Quest 3 mixed reality flagship. Until November 12, 2025, when Valve suddenly dropped three bombs: Steam Machine console, Steam Controller, and Steam Frame VR headset. Early 2026 launch, pricing unannounced, but estimated at $499-599 for 256GB and $699-799 for 1TB.

Steam Frame is not a simple Index upgrade, but a “streaming-first” standalone VR headset. It wants to answer one question: if a VR headset can both play mobile-level VR independently and wirelessly stream PC AAA titles, why bother with cables?

Product Overview: The VR Piece of Steam’s Ecosystem Puzzle

Valve’s hardware strategy is clear: Steam Deck handles mobile scenarios, Steam Machine handles living room scenarios, and Steam Frame handles immersive scenarios. All three devices share the SteamOS system, the same game library, and the same microSD card. Your saves, settings, and progress seamlessly switch between the three devices.

Steam Frame’s core positioning is “streaming-first.” It packs a Snapdragon 8 Gen 3 processor and 16GB memory, capable of running some VR games and Android apps independently, but the real selling point is the 6GHz wireless adapter—plug it into your PC, and the headset directly streams PC VR content with low latency. Dual-radio architecture: one handles audio-visual streaming, one handles WiFi networking, without interference.

Specifications: Valve’s Spec Monster Sincerity

Spec	Details
Processor	Snapdragon 8 Gen 3 (4nm)
Memory	16GB LPDDR5X
Storage	256GB/1TB + microSD
Display	Dual 2160×2160 LCD
Refresh Rate	72/90/120Hz, experimental 144Hz
Field of View	110°
Optics	Pancake lenses
Tracking	Inside-out 6DoF (4 monochrome cameras)
Eye Tracking	Yes, Foveated Streaming
Wireless Streaming	6GHz dedicated adapter, dual-radio
Weight	~440g (with battery)
Battery	Rear-mounted design, front-rear weight balance
System	SteamOS VR edition
Compatibility	SteamVR, OpenXR, Android apps

Data source: Valve official launch materials, Wareable, Engadget, VideoCardz

The Snapdragon 8 Gen 3 is a 2024 flagship phone chip, with performance near phone-level rather than PC-level. This means Steam Frame’s standalone capabilities are limited—running Half-Life: Alyx smoothly? Impossible. But paired with 6GHz wireless streaming, a PC’s RTX 4090 can be “lent” to the headset, which is Valve’s true intention.

Dual 2160×2160 resolution per eye, totaling approximately 4.3K, slightly higher than Quest 3’s 2064×2208 but far below Vision Pro’s 4K Micro-OLED. Pancake lenses make the body thinner, but optical efficiency is inferior to traditional Fresnel lenses, and brightness may be limited. The 110° field of view is moderate—narrower than Index’s 130° but on par with Quest 3’s 110°.

Foveated Streaming is Steam Frame’s technical highlight. Through eye tracking, the system only renders high-resolution images in your gaze area, reducing resolution in peripheral areas to save substantial bandwidth. Official claims suggest bandwidth efficiency improvements of over 10x—meaning equivalent quality with less wireless streaming latency and compression artifacts.

Valve Steam ecosystem product lineup display

Deep Analysis: Where Steam Frame’s Chances Lie

Valve has three clear competitive advantages with Steam Frame:

First, Steam library ecosystem moat. Steam has over 70,000 games, including more than 5,000 VR titles. The Quest 3 store has only a few hundred VR games, and Meta’s strict content review means many PC VR classics (like Half-Life: Alyx, Beat Saber) have no Quest versions. Steam Frame directly compatible with the entire SteamVR library—an advantage Meta cannot replicate.

Second, open platform positioning. Meta Quest is a closed ecosystem; apps must pass Meta store review. Steam Frame runs SteamOS, supports OpenXR standards, and developers can freely publish content. Valve even provides CAD drawings and circuit specifications, letting users 3D print replacement face cushions and head straps—this “modular” thinking is unique in the VR industry.

Third, PC streaming experience optimization. Existing wireless streaming solutions (like Virtual Desktop, Air Link) rely on home WiFi, with unstable latency and image quality. Steam Frame’s 6GHz dedicated adapter bypasses the router, directly establishing a private PC-to-headset link, paired with Foveated Streaming, theoretically achieving near-wired connection experience.

But Steam Frame’s challenges are equally severe:

Insufficient standalone performance. The Snapdragon 8 Gen 3 can probably run Resident Evil 4 VR, but Half-Life: Alyx and Microsoft Flight Simulator must rely on PC streaming. This means users without high-end PCs can only play “mobile-level” VR, significantly compromising the experience.

Weight and comfort. At 440g including battery, lighter than Quest 3’s 515g but much lighter than Index’s 809g. The rear-mounted battery design helps front-rear weight balance, but comfort during extended wear (over 1 hour) remains to be verified. Pancake lenses are thin, but edge clarity may degrade.

Price competitiveness. Estimated $499-799 pricing offers no advantage over Quest 3 ($499) and Quest 3S ($299). Vision Pro is expensive ($3,499) but positioned completely differently. Steam Frame needs to convince users: is spending $200-300 more for the Steam library and open ecosystem worth it?

Content moderation gray area. Valve’s open strategy means adult, violent, and extreme content may more easily enter the platform. While Steam PC already has mature content rating systems, VR’s immersion makes sensitive content more impactful. How Valve balances “openness” and “compliance” is a long-term challenge.

Comparison: Steam Frame vs Meta Quest 3 vs Apple Vision Pro

Feature	Steam Frame	Meta Quest 3	Apple Vision Pro
Price	$499-799 (estimated)	$499 (128GB)	$3,499
Weight	~440g	~515g	600-650g
Resolution	2160×2160 per eye	2064×2208 per eye	4K Micro-OLED per eye
Refresh Rate	72-144Hz	72-120Hz	90-100Hz
Field of View	110°	110°	100°
Tracking	Inside-out 6DoF	Inside-out 6DoF	Inside-out 6DoF
Eye Tracking	Yes (Foveated Streaming)	No	Yes (Foveated Rendering)
Wireless Streaming	6GHz dedicated adapter	WiFi (Air Link)	Only Mac/iPhone mirroring
Standalone Performance	Snapdragon 8 Gen 3	Snapdragon XR2 Gen 2	M2+R1 (extremely powerful)
Game Library	SteamVR (5000+ VR games)	Meta Store (few hundred)	Very few native VR games
Ecosystem Openness	Fully open	Closed review	Closed review
Mixed Reality	Monochrome passthrough (no color)	Full-color passthrough	Full-color passthrough

Steam Frame’s differentiation is clear: it’s not the strongest standalone device (Vision Pro), nor the cheapest (Quest 3S), but the “best wireless companion for PC VR players.” If you already have a high-end PC and Steam library, Steam Frame is an upgrade replacement for Quest 3; if you only want to play standalone VR games, Quest 3 offers better value; if you pursue spatial computing and productivity, Vision Pro is another dimension.

Pros and Cons

Pros	Cons
Full SteamVR library compatibility, unbeatable game resources	Limited standalone performance, heavy games depend on PC streaming
6GHz dedicated streaming, theoretically better latency and quality	440g weight comfort during extended wear to be verified
Foveated Streaming saves bandwidth, improves experience	Higher price than Quest 3, value advantage not obvious
Fully open ecosystem, high developer freedom	No color passthrough, mixed reality weaker than Quest 3
Modular design, customizable accessories	Content moderation gray area, sensitive content risks
Shared ecosystem with Steam Deck/Machine	Shipping date not final, potential delays

Who Should Buy

Recommended for:

Players with high-end PCs and SteamVR game libraries
VR enthusiasts tired of Quest’s closed ecosystem, seeking open platforms
Business travelers needing portable VR large screens
Valve Index legacy users wanting wireless upgrade without abandoning Steam library

Not recommended for:

Users without high-end PCs, only wanting standalone VR games (Quest 3 more suitable)
Budget-constrained users—$299 Quest 3S is sufficient
Users pursuing mixed reality and productivity scenarios (Vision Pro or Quest 3 better)
Users demanding brand stability—Valve hardware often delays and stocks out

FAQ

Q: Does Steam Frame require Steam Machine?

A: No. Any PC supporting SteamVR can pair with the 6GHz adapter for wireless streaming. Steam Machine is just Valve’s recommended “living room companion.”

Q: Does Foveated Streaming affect image quality?

A: Theoretically no. Eye tracking ensures your gaze area is always full resolution, with only peripheral areas reduced—human peripheral vision naturally has low resolution, so no noticeable difference is perceived.

Q: Does it support prescription lenses?

A: Valve has not announced, but Index supports custom lens inserts, so Steam Frame likely will have similar solutions.

Q: How long is the battery life?

A: Official data not announced. Referencing Quest 3’s 2-3 hours, Steam Frame’s 440g weight may allow a larger battery, but 6GHz streaming consumes significant power, so actual battery life may be similar.

Conclusion

Steam Frame is Valve’s “belated response” to the VR industry. Six years ago, Index defined the ceiling of PC VR; in six years, Meta’s Quest series captured the standalone VR market. Steam Frame is not Index’s successor, but a new species—it acknowledges standalone VR’s performance limitations, instead using wireless streaming to “borrow” PC computing power to the headset, while preserving the open advantages of the Steam ecosystem.

This strategy is smart and risky. Smart because it avoids head-to-head competition with Quest 3 on standalone performance, instead building differentiation through the Steam library and open ecosystem; risky because whether wireless streaming can truly approach wired experience, and whether $499-799 pricing can convince enough PC VR players to upgrade.

2026 will be a pivotal year for VR hardware: Meta Quest 4, Apple Vision Pro 2, and Sony PS VR2 successor may all debut. Whether Steam Frame can stand out in this melee depends on three factors: whether streaming latency is low enough, whether pricing is competitive enough, and whether Valve’s production can keep up (referencing Steam Deck’s long-term stock shortages).

2026-06-06

KANNAN K2 AI Smart Glasses Review: 25.8 Grams of First-Person AI Capture

Introduction: When AI Glasses Become Light Enough to Forget

On June 5, Kannan Tech released the K2 smart glasses in Hangzhou. Twenty-five point eight grams total weight, 7.1mm temple thickness at the thinnest point, European-imported ultra-light nylon material—these three numbers alone tell me this company is not playing with concepts but with materials science. Priced at 2350 yuan, with a 32MP camera, 4K 30FPS recording, EIS plus AI dual stabilization, and Kiro OS built specifically for AI glasses.

Honestly, the AI glasses track is already insanely competitive. Meta Ray-Ban, Vidda G11, RayNeo V3—each finding their own differentiation. The K2’s approach is crystal clear: no AR display, no complex interactions, just focus on doing one thing well—letting you capture first-person life with the lightest weight and most natural approach. This sounds simple, but achieving 25.8 grams, lighter than Vidda G11’s 26.5g and nearly half the weight of Meta Ray-Ban’s ~49g, is itself a demonstration of technical strength.

Product Overview: Kannan Tech’s AI Glasses Ambition

Kannan Tech is an AI glasses startup headquartered in Hangzhou, China, with unclear team background but solid product definition capabilities. The K2’s positioning is very clear: first-person AI capture glasses for general consumers, not a geek toy, not a business tool, but a “life recording” carry-everywhere device.

The K2’s design language is “invisible.” Five frame styles available, from classic black to fashionable titanium, aiming to make it unnoticeable that you are wearing smart glasses. European-imported ultra-light nylon material, combined with internal structure optimization, compresses total weight to 25.8 grams—approaching regular optical glasses weight, with no obvious pressure during extended wear.

Specifications: How Much Tech Fits in 25.8 Grams

Spec	Details
Weight	25.8g (total)
Temple Thickness	7.1mm (thinnest point)
Material	European-imported ultra-light nylon
Camera	32MP, f/2.4 aperture
Video Recording	4K 30FPS
Stabilization	EIS + AI dual stabilization
Photo Features	Live Photo support
Operating System	Kiro OS (built for AI glasses)
Processor	WQ7036 smart audio main control chip
AI Assistant	Xiaonan
Interaction	Minimalist gestures + voice
Special Features	AI flash memo, message sharing
Battery	60 min recording / 7 days standby / 1.5 days typical use
Price	2350 yuan

Data source: Kannan Tech official launch materials

Twenty-five point eight grams—what does this mean? Regular prescription glasses typically weigh 20-30g, so the K2 falls completely within the “daily wearable” range. The 7.1mm temple thickness at the thinnest point makes it almost indistinguishable from regular glasses from the side. Kannan Tech invested heavily in material selection—European-imported ultra-light nylon, lighter, tougher, and more wear-resistant than traditional TR90 material.

The 32MP camera is high-end among AI glasses. Meta Ray-Ban’s camera is 12MP, RayNeo V3’s Sony IMX681 is around 12MP equivalent, so the K2’s 32MP clearly leads in hardware specs. The f/2.4 aperture is relatively large for glasses cameras, allowing more light intake in low-light environments, theoretically improving night shooting performance.

EIS plus AI dual stabilization is a practical configuration. EIS (Electronic Image Stabilization) compensates hand shake through algorithms, while AI stabilization further optimizes through scene recognition and motion prediction. For walking, cycling, handheld shooting scenarios, dual stabilization significantly improves frame usability.

Hands-On Experience: What Can First-Person Capture Actually Do?

The K2’s core usage scenarios fall into three categories: life recording, content creation, and social sharing.

Life Recording: Playing with kids, pet interactions, travel experiences—these scenarios captured with traditional cameras or phones always have a “third-person perspective” sense of distance. The K2’s first-person capture shows exactly what the glasses see, without deliberate composition or the awkwardness of holding up a phone. The Live Photo feature makes photos “come alive,” recording 1.5 seconds of motion before and after the shutter press, giving ordinary photos more storytelling power.

Content Creation: Vlog creators, food bloggers, exploration influencers—the K2’s 4K 30FPS recording quality is sufficient for short video platform publishing. The AI flash memo feature lets you quickly trigger “remember this” by voice when seeing exciting moments; the AI automatically captures surrounding clips and categorizes storage, saving post-production footage review time. The message sharing feature can generate 15-second short videos from highlight clips, sharing directly to WeChat, TikTok, and Xiaohongshu with one click.

Social Sharing: Friend gatherings, family dinners, outdoor activities—the K2’s minimalist gesture interaction (tap temple to shoot, slide to switch modes) makes capturing natural and non-awkward. You don’t need to pull out your phone, open the camera, and hold it up—the entire process completes within 2 seconds without interrupting the present moment.

But the K2 also has obvious limitations:

No AR Display: The K2 has no built-in display; all operation feedback comes through the mobile app or voice announcements. This means you cannot see shooting previews on the glasses, only “blind shooting” and reviewing through the phone later. For composition-critical scenarios, this is a pain point.

Battery Bottleneck: Sixty minutes of recording battery life is mid-range among AI glasses, but heavy use (continuous Vlog shooting) requires carrying a charging case. Seven days of standby sounds long, but the “typical daily use” 1.5-day definition is vague—if you shoot 20 minutes of video daily and use the AI assistant for 30 minutes, actual battery life may be less than a day.

Limited AI Assistant Depth: The Xiaonan assistant supports basic functions like voice weather queries, setting reminders, and playing music, but compared to Siri or Xiao Ai on phones, feature coverage is clearly insufficient. The K2’s AI capabilities focus on “shooting assistance” rather than “general intelligence”—clear positioning but a visible ceiling.

Comparison: K2 vs Meta Ray-Ban vs RayNeo V3

Feature	KANNAN K2	Meta Ray-Ban	RayNeo V3
Weight	25.8g	~49g	~39g
Price	2350 yuan	~$299 (~2150 yuan)	973 yuan (after price drop)
Camera	32MP	12MP	Sony IMX681 (~12MP)
Video	4K 30FPS	1080P	1080P
Stabilization	EIS + AI dual	Basic EIS	Basic EIS
AI Features	Flash memo, message sharing	Social sharing, Meta AI	200+ scenario AI
Audio	Open-ear directional	Open-ear	Open-ear
Battery	60 min recording	~36 hours combined	4-6 hours
Ecosystem	Standalone app	Meta ecosystem	RayNeo ecosystem
Positioning	First-person capture	Social + AI	Exploratory AI

Three glasses, three paths: Meta focuses on social ecosystem, RayNeo on AI exploration, K2 on capture quality. The K2’s 32MP camera and 4K recording clearly lead in hardware, but software ecosystem and AI depth lag behind Meta and RayNeo. If you pursue “shooting clearly, recording steadily, sharing fast,” the K2 is the best choice; if you need “all-powerful AI assistant, seamless social integration,” Meta is more suitable; if you want to “try AI features on a budget,” the RayNeo V3 offers the best value.

KANNAN K2 battery life promotional feature

Pros and Cons

Pros	Cons
25.8g extreme lightweight, top-tier wearing comfort	No AR display, cannot preview shots in real-time
32MP + 4K 30FPS, leading hardware specs	60-minute recording battery, heavy use needs charging case
EIS + AI dual stabilization, good frame stability	Limited AI assistant depth, inferior to phone voice assistants
Kiro OS optimized for AI glasses, smooth interaction	Low brand recognition, after-sales and accessory ecosystem to be built
Minimalist gesture interaction, low learning curve	2350 yuan mid-range pricing, value inferior to RayNeo V3
Live Photo and AI flash memo, more storytelling	Chinese-only interaction, limited for overseas users

Who Should Buy

Recommended for:

Users pursuing extreme lightweight, long-term wear without nose bridge pressure
Vlog creators and short video bloggers needing high-quality first-person footage
Parents and pet owners wanting to record daily interactions without phone interruption
Travel enthusiasts needing portable, discreet capture devices

Not recommended for:

Users needing AR display and real-time information overlay (choose RayNeo or Meta)
Users pursuing comprehensive AI features and rich ecosystem (choose Meta Ray-Ban)
Budget-constrained users only accepting under 1000 yuan (choose RayNeo V3)
Users demanding brand after-sales and accessory ecosystem (wait for K2 ecosystem maturity)

FAQ

Q: Does the K2 support prescription lens customization?

A: Yes. The K2 provides prescription lens customization services, available at purchase or through designated optical shops.

Q: How is the photo quality from the 32MP camera?

A: Hardware specs lead, but actual image quality is limited by optical design constrained by glasses form factor. Excellent in daylight with sufficient light; noise control is average in low-light, inferior to phone main cameras.

Q: Does the AI flash memo feature require internet?

A: Basic flash memo functions support local processing, but generating message sharing and cloud storage require internet. Recommend using in Wi-Fi environments for optimal experience.

Q: What is the K2’s water resistance?

A: Daily splash-proof level, handling sweat and light rain. Not recommended for swimming, showering, or prolonged immersion.

Conclusion

The KANNAN K2 is a product that “does one thing to the extreme.” In an AI glasses track where everyone pursues “all-in-one,” the K2 chooses subtraction—removing AR display, simplifying the AI assistant, focusing on capture experience, establishing clear differentiation through 25.8 grams of extreme lightweight and 32MP plus 4K hardware specs.

This strategy carries both risks and opportunities. The risk: the ultimate form of AI glasses may be “all-in-one terminals,” and the K2’s single-function positioning may limit long-term market space. The opportunity: before all-in-one terminals mature, “professional tools” often gain more loyal user bases—just as DSLR cameras were never replaced by phones, first-person capture devices may also become an independent category.

At 2350 yuan, pricing sits in the mid-range—slightly more expensive than Meta Ray-Ban but with clearly higher hardware specs. For users pursuing capture quality and wearing comfort, the K2 is worth considering. But if you prioritize AI ecosystem and social integration, Meta Ray-Ban remains the safer choice. The K2’s true test lies in whether Kannan Tech can continuously iterate Kiro OS, enrich AI features, and build accessory and after-sales ecosystems—these soft capabilities are the key determining whether an AI glasses product can evolve from “niche gem” to “mass-market hit.”

2026-06-06

WIKO Xingzai Unboxing: Do Young People Need This?

Introduction: When AI Evolves from Tool to Pet

WIKO Xingzai AI companion toy promotional poster

At the Huawei Nova 16 launch event on May 30, alongside phones, a fluffy little creature emerged—Xingzai. It is not a phone accessory, not a smart speaker, but an AI companion toy that hangs on your bag. It turns its head toward you when you speak, shrinks shyly when you pet its forehead, and shakes excitedly when you sway it. Priced at 499 yuan, powered by Huawei Xiaoyi large model, with Wi-Fi plus 4G dual-network independent operation.

Honestly, my first reaction seeing the product images was: is this not just an upgraded electronic pet? But after getting hands-on, I discovered WIKO aims far beyond that. Xingzai’s core selling point is not “can talk” but “reacts”—every head turn, every light change, every body tremor simulates real pet feedback logic.

Product Overview: A New Species in Huawei’s Ecosystem

First, clarifying relationships: WIKO is not a Huawei subsidiary but a deep ecosystem partner. Huawei provides Xiaoyi large model and HarmonyOS ecosystem connectivity; WIKO handles R&D, design, production, and sales. Xingzai is WIKO’s first AI companion product built on Huawei technology, with the nova planet IP licensed from Huawei’s youth-focused nova brand.

Xingzai measures 110x110x125mm and weighs approximately 260g, similar to a smartphone. The magnetic design attaches directly to backpacks or purses, becoming a fashion accessory on the go. This positioning is clever—traditional AI toys are mostly bulky and home-bound, while Xingzai breaks scenario limitations entirely.

Specifications: How Much Tech Fits in This Small Body

Spec	Details
Dimensions	110x110x125mm
Weight	~260g
Attachment	Magnetic design
AI Model	Huawei Xiaoyi large model
Networks	Wi-Fi + 4G dual
Battery	1900mAh
Battery Life	6 hours continuous dialogue / 48 hours standby
Interactions	Voice + touch + shake + sound-following head turn
Memory	Exclusive diary memory, intimacy unlocks
Price	499 yuan

Data source: WIKO official launch materials

Huawei fluffy AI pet toy handheld display

Hands-On Experience: Does It Actually Touch Your Heart

Sound-Following Head Turn: This is Xingzai’s most moving feature. Speak from the left, its head turns left, starry eyes lighting up, simulating “I am listening to you attentively.” The psychological logic is precise—humans have innate affinity for “being watched,” and head-turning delivers stronger emotional impact than mere voice replies.

Touch Interaction: Petting the forehead triggers shyness (softened lights + slight body shrink), while prolonged petting triggers (flashing lights + gentle trembling). These reactions are not random but based on Xiaoyi large model’s learning of interaction frequency and duration—the more you interact, the more it understands you.

Shake Reaction: A gentle shake makes Xingzai tremble excitedly, with starry eyes flashing rapidly, simulating a pet being amused. This scenario shines at friend gatherings, maximizing social appeal.

Voice Dialogue: Powered by Xiaoyi large model, natural language understanding is solid. It chats about daily life, tells jokes, offers advice, and adjusts responses based on your mood. When you are down, it comforts you; when happy, it celebrates with you. The exclusive diary memory system records your conversations, forming “shared memories,” with more interaction modes unlocking as intimacy grows.

Independent Operation: Wi-Fi plus 4G dual-network support means no phone connection needed outdoors. Whether on the subway, at a cafe, or camping, Xingzai works normally. Traditional AI toys become mere plush fabric without phones; Xingzai solves this pain point.

AI companion pet toy physical product display

Comparison: WIKO Xingzai vs Traditional AI Toys vs LOVOT

Feature	WIKO Xingzai	Traditional AI Story Machine	Japan LOVOT
Price	499 yuan	200-500 yuan	~3000 yuan + monthly fee
Portability	Magnetic bag charm, portable	Home use only	Home use only
Dynamic Interaction	Head turn + tremor + lights	None	Movement + expressions + body heat
AI Model	Huawei Xiaoyi large model	Basic voice chip	Self-developed emotional AI
Networks	Wi-Fi + 4G independent	Requires phone/Wi-Fi	Requires Wi-Fi
Positioning	Youth fashion accessory	Children’s education	Premium emotional companion

Xingzai’s differentiation is clear: more dynamic interaction and independent networks than traditional AI toys, an order of magnitude cheaper than LOVOT. At 499 yuan for an emotionally responsive, go-anywhere AI buddy, the barrier is low for young people.

Pros and Cons

Pros	Cons
Magnetic design, truly portable companionship	260g is slightly heavy on bags
Multi-modal interaction, genuine emotional feedback	6-hour battery, heavy use needs daily charging
Huawei Xiaoyi large model, natural conversations	Plush material not dirt-resistant, outdoor use requires care
499 yuan friendly price, low trial cost	Limited functional depth, unsuitable for deep emotional dependence
Wi-Fi + 4G independent, no phone tethering	Currently supports Chinese interaction only

Who Should Buy

Recommended for:

Young female users who love cute accessories
Urban youth living alone needing emotional companionship
Students curious about AI toys on limited budgets
Frequent outdoor activity users needing portable companions

Not recommended for:

Pragmatists indifferent to plush toys
Users needing deep emotional anchors (professional counseling recommended)
Consumers seeking premium robot experiences

FAQ

Q: Does Xingzai require a Huawei phone? A: No. Xingzai supports independent Wi-Fi and 4G operation; any brand phone can pair, though Huawei/HarmonyOS users get the most complete experience.

Q: Is the diary memory function secure? A: Dialogue data is stored locally and on Huawei cloud, following Huawei privacy protection policies. Users can delete memory records anytime.

Q: Is it washable? A: The plush cover is detachable for cleaning; internal electronics must not be submerged. Wipe the shell regularly with a damp cloth.

Conclusion

WIKO Xingzai is not a technological revolution but a scenario innovation. It transforms AI companionship from “toys left at home” to “buddies taken outdoors,” lowering the experience barrier for AI emotional products at a friendly 499 yuan price. With Huawei Xiaoyi large model and dual-network independence, it has no direct competitors at this price point.

But Xingzai faces a fundamental question: what is the ultimate value of AI companion toys? Replacing real pets, or serving as temporary patches for loneliness? If Xingzai makes you smile while waiting for the subway or gives you someone to talk to during late work nights, then 499 yuan is worth it. But if you expect it to become a true emotional pillar, you may need to wait for AI to evolve a few more generations.

2026-06-05