Somni turns bedside sensing into an overnight intervention

Somni, Fullive.ai's approaching-mass-production sleep device, turns bedside sensing into a proposed overnight loop of detection and intervention. According to geekpark, millimeter-wave radar tracks breathing and heart rate. A microphone array tracks movement and snoring. Environmental sensors measure temperature. They also measure humidity. Light is monitored as well. Its intended sequence is explicit: sense a user's state, then choose a strategy. It can deliver sound. It can also use light or scent, then use the response to refine the next intervention. Fullive.ai calls that logic "state -> intervention -> response."

The emerging question is what makes such a loop worth trusting. A report describes Lingluo Technology's ankle-worn device as something used briefly before sleep, then removed-a sharply different bargain from an all-night bedside system.

A sleep score is becoming a control system

Somni, Fullive.ai's sleep hardware device for high-pressure knowledge workers, frames the sleep score as the start of a decision rather than the final report. Approaching mass production, it is built to run beside the bed instead of asking the user to wear it overnight.

Its sensing stack combines millimeter-wave radar, a microphone array, and environmental sensors. It detects breathing and heart rate, then tracks body movement and snoring; it also reads bedroom temperature. Humidity and light are monitored as well. Those inputs feed a closed loop: sense a user's state, determine a strategy, deliver an intervention, then iteratively optimize it. Fullive.ai describes the underlying logic as state -> intervention -> response, with the resulting record called response data.

The intended interventions include sound. They also include light and scent. Somni is designed to help users fall asleep faster, reduce nighttime awakenings, and wake naturally, making its value depend on what happens after detection.

Lingluo Technology presents a more bounded pre-sleep model. Its device is worn at the ankle for 15 minutes before sleep, then removed. The company claims its Temporal Interference technology enables noninvasive targeting of deep neural tissue. That approach places stimulation before bed rather than maintaining a bedside sensing-and-response loop through the night.

Shenyu Technology pushes the control-system idea toward a medical device. Its first medical product is a closed-loop temporal interference electrical stimulation device, designed for non-invasive, precise stimulation of deep brain regions. Multimodal monitoring feedback is intended to dynamically optimize stimulation plans. According to the report, the shared pattern is feedback guiding intervention, but the practical form differs: Somni changes the bedroom environment, while Lingluo and Shenyu center electrical stimulation.

How three Chinese sleep and brain-health systems approach the sensing-intervention loop

The intervention layer divides these products

Somni's intervention happens at the bedside rather than on the body. According to geekpark, the screen-free device sits beside the bed and does not require the user to wear it. Fullive.ai says it combines sound with light. Scent is also part of the system. It is intended to speed sleep onset. It also limits nighttime awakenings and supports a natural morning wake-up.

That design makes the whole night the intervention window. Fullive.ai describes each use as a small experiment: Somni observes a user's state, changes the environment, then assesses subsequent physiological changes. The premise is adaptive rather than fixed. A bedside system can keep adjusting without asking someone to put on hardware while trying to sleep. The user need not charge it then or remove it.

Lingluo Technology takes a narrower, more direct route. Its device is worn at the ankle for 15 minutes before sleep, according to the outlet, then removed. The aim is to produce relaxation and drowsiness before the user gets into bed, rather than alter bedroom conditions through the night.

The contrast is about where intervention belongs. Somni treats sleep as a period for continuing adjustment; Lingluo concentrates its effect in the transition into sleep. The outlet reports Lingluo's claim that nearly all users in its three-month offline experience testing felt obvious drowsiness and deep relaxation shortly after a 15-minute trial. The company also claims its first product sold out within 48 hours of its initial sale in July and expects its second-generation product to launch overseas within the next six months.

Somni is earlier in a different sense. geekpark reports small-batch participant experiments, alongside pursuit of mass production this year and larger-scale clinical validation. The intervention layer therefore separates an all-night ambient system from a brief pre-bed wearable ritual.

Wear time changes the adoption bargain

Somni makes wear time disappear by moving sensing to the bedside. According to geekpark, its unit combines millimeter-wave radar with a microphone array. It also uses environmental sensors to track breathing and heart rate. It tracks movement and snoring as well as temperature, humidity, and light.

It has no screen and requires no device on the body.

Placed beside the bed, Somni is designed to sense a user's state. It can then select a strategy, deliver an intervention, and refine that intervention over time.

That changes the adoption bargain. A person can leave Somni in place rather than remember to attach a sensor as part of bedtime. The person does not need to charge it or remove it. The trade-off is that a bedside system must infer the body from a distance, while also accounting for conditions in the room. Its convenience is therefore part of the product design, not a secondary benefit of the sensing hardware.

Lingluo Technology makes a different compromise. Its ankle device is intended to be worn for 15 minutes before sleep and then removed, according to a cited report. The limited-scale user experience testing that began in early 2026 included offline activities lasting three months. That short wear window may reduce overnight burden, but it also asks users to complete a deliberate pre-sleep step.

Shenyu Technology takes the more intimate route: wearable brain-computer interface products for everyday brain-function monitoring and intervention. Its temporal-interference electrical-stimulation device is designed to non-invasively target deep brain regions. It uses multimodal monitoring feedback to adjust stimulation plans. A cited report says that the company also seeks large-scale real-world EEG and multimodal data through consumer use to improve personalized models and later product iterations.

The relevant specification is not simply what each sensor can detect. It is the routine each product asks a person to accept.

Early user response is not clinical proof

Somni's small-batch participant experiments are an early development step, not a clinical verdict. According to geekpark, the company is pursuing mass production this year while seeking larger-scale clinical validation. That sequence matters: a device can show promise in a limited group yet still need broader evidence on outcomes and safety, including sustained use.

Lingluo Technology offers a different kind of early signal. A cited report says that its limited-scale user experience testing began in early 2026 and included offline activities lasting three months. The company claims that nearly all participants felt clear drowsiness and deep relaxation shortly after a 15-minute trial, and says its first product sold out within 48 hours of its initial sale in July.

Those claims may indicate curiosity and immediate subjective response. They do not establish treatment effectiveness.

For Fullive.ai, the larger clinical validation sought for Somni is therefore the more consequential threshold. It would move the discussion beyond whether participants notice a short-term effect toward whether an intervention reliably improves relevant sleep outcomes under appropriate study conditions.

Shenyu Technology's route makes the distinction sharper. It has pursued serious medical care alongside consumer health since its founding, and its first medical product is a closed-loop temporal interference electrical stimulation device. A cited report says funding is intended in part for clinical validation and registration conversion of medical-grade products for auxiliary depression diagnosis and neuromodulation. The team also claims industrialization leadership in China, but that positioning does not replace validation or registration.

The ledger provides no comparative evidence against CBT-I, melatonin, prescription drugs, meditation apps, or ordinary sleep-hygiene changes.

Choosing among sensing, coaching and neuromodulation tools

• You want overnight sleep monitoring without wearing a device. Consider Somni's bedside approach: it is designed to use millimeter-wave radar, a microphone array and environmental sensors, with no screen or required wearable. Its intended loop is sensing, strategy selection, sound/light/scent intervention and iterative optimization. However, it is still pursuing mass production and larger-scale clinical validation, so treat its sleep-improvement claims as product claims rather than established clinical evidence.
• Your immediate problem is falling asleep under stress, and you are willing to use a wearable briefly before bed. Lingluo's ankle-worn device is the more targeted option in the source material: it is intended to be worn for 15 minutes before sleep and then removed. The company focuses on neural intervention and claims its Temporal Interference technology enables noninvasive deep neural targeting. Its reported drowsiness results come from limited-scale user experience testing, so do not equate them with independently described clinical validation.
• You are considering a device that electrically stimulates the brain for sleep, depression or another brain-health condition. Use a much higher safety threshold. Shenyu is developing EEG-based monitoring and precision neuromodulation, including electrical stimulation and temporal interference stimulation, and its first medical product is a closed-loop temporal interference electrical stimulation device. The company is also funding clinical validation and registration conversion of medical-grade products; the sources do not establish availability, approval status or outcomes for consumer use.
• You need to make sense of symptoms, medication changes, sleep, stress and fatigue over months before a clinical appointment. A longitudinal record tool such as Juno is the better fit than a sleep intervention device. It is designed to organize text or voice conversations, body metrics, medical history and wearable data into a long-term profile, identify possible patterns, and generate a pre-appointment PDF. Its pattern-finding is explicitly not a diagnosis or medical causal finding, and it is not presented as a replacement for professional doctors.

Treat bedroom data and stimulation as trust questions

Bedroom products earn trust by making the loop inspectable. Ask what intervention the system delivers, which physiological response changes that intervention, and what evidence supports the claimed benefit. A sleep score alone cannot answer those questions.

The evidence bar rises when sensing leads to stimulation.

According to geekpark, Fullive.ai says Somni stores raw radar waveforms locally first. Information sent to the cloud is encrypted and minimized, while uploads of raw material capable of reconstructing personal privacy are avoided where possible. That local-first approach is a meaningful design choice, but it leaves practical questions: what remains on the device, who can access it, and how users can control or remove it.

Somni's stated Physiological World Model is meant to learn how physiological states change with environment and behavior. It also accounts for intake. It tracks rhythms and interventions. That ambition makes data rights central, because the model's value depends on linking intimate overnight signals to actions and outcomes. Users should ask what data are retained and how those links are governed.

Shenyu Technology shows the adjacent consumer-to-medical path. A report says that it uses wearable brain-computer interface products for everyday monitoring and intervention, while developing EEG-based auxiliary diagnosis and precision neuromodulation involving electrical stimulation. It also covers temporal-interference and acoustic-electrical stimulation. Its planned funding covers consumer mass production, overseas expansion, and clinical validation and registration conversion for medical-grade products. Those plans do not settle safety or overseas regulatory status. They make the unanswered questions clearer: which uses remain consumer tools, which require medical oversight, and what validation supports each intervention.

For neurostimulation or brain-computer interface tools, separate a product roadmap from proof that belongs in a bedroom routine. Shenyu Technology is developing combined focused-ultrasound and electrical stimulation, while its founder describes a plan to integrate AI with brain-computer interfaces for long-term public brain-health use. Those are ambitious directions, not a substitute for evaluating the specific intervention a device delivers or the feedback it uses to adapt.

Ask vendors to distinguish claims from demonstrated use. According to a Chinese technology outlet, Shenyu says it is the first team in China systematically pursuing industrialization of this combined acoustic-electrical approach. Examine what that distinction means for the actual product, rather than treating investor participation or a team's backgrounds at United Imaging, Mindray, ByteDance, and Tencent as evidence that an overnight intervention is ready to trust.

For readers outside China

• Availability: Somni is approaching mass production, while Lingluo's first AI sleep wearable device officially went on sale in China in July 2026. Lingluo expects to officially launch its second-generation product overseas within the next six months. The source material does not say which overseas markets will be included, whether Somni is available outside China, or where Shenyu's products can be bought. Juno was publicly launched in October 2025, but the sources do not say in which countries it is available.
• Pricing: Not disclosed in sources.
• Closest Western equivalents: A contactless bedside sleep tracker combined with a smart sleep-environment device; A short-session wearable sleep-onset intervention device; A chronic-condition symptom journal that turns ongoing records into a clinician-ready appointment summary; A closed-loop neuromodulation and brain-monitoring platform
• Data residency: Fullive.ai says Somni initially stores raw radar waveforms locally and uses encrypted, minimized cloud uploads, avoiding uploads of raw data that could reconstruct personal privacy where possible. This is a company statement, not a disclosed data-residency policy. The source material does not cover server locations, retention periods, deletion controls, cross-border transfers, or Juno, Lingluo and Shenyu data-handling practices.

Sources

• 36kr 李泽湘孵化的睡眠AI硬件公司获新融资，深高新投领投数千万元｜硬氪首发 https://36kr.com/p/3954623000952197
• geekpark 成立不到一年连融三轮，专治失眠的 AI 床头灯「火」了 https://geekpark.net/news/369768
• 36kr 超千万元天使轮融资落定！ 深圳神奕科技有限公司：AI融合BCI，让脑机智能普惠大众 https://36kr.com/p/3963142866369923
• woshipm 两个被慢性病折磨多年的年轻人，做了个"记住你全部病史"的 AI，8个月15万下载 https://woshipm.com/chuangye/6453455.html