Sensory Launches WASM-Powered Browser Demos Showcasing Wake Words, Speech-to-Text & NLU, and Phrase Spotted Commands

Santa Clara, Calif. – May 27, 2026 – Sensory, a leading provider of high-accuracy, low-power voice and vision AI, today announced the availability of three new interactive WebAssembly (WASM)–powered product demos that allow developers and product teams to experience Sensory’s on-device voice technology directly in a standard web browser. The demos showcase real-time wake word detection, speech-to-text (STT) and natural language understanding (NLU), and phrase spotted voice commands, all running at the edge for fast, reliable and private voice interactions.

Available now at Sensory.com, the browser-based experiences are designed for teams building wearables, automotive systems, smart home devices, healthcare products, and connected devices that demand low-latency, always-available voice interfaces without the privacy risks and dependency of cloud-only architectures. By compiling Sensory’s proven embedded models to WASM, the demos also show how the same technology can be embedded into virtually any website or web application, enabling brands to add voice AI to product pages, support flows and interactive online experiences.

See Sensory’s voice AI in the browser

The new demos include three focused, interactive experiences that launch instantly in the browser and require no hardware, downloads or plug-ins.

Wake Words – A real-time wake word demo lets users experience Sensory’s ultra-low power wake word technology across popular assistants such as “hey siri,” “ok google,” “hey meta,” “alexa” and “hey copilot.” The demo highlights how Sensory wake words run entirely on-device, are optimized for small footprints and are engineered for robust performance across noisy, real-world environments in any language.

Speech-to-Text & NLU – A live transcription demo showcases Sensory’s embedded Speech-to-Text and NLU engine, delivering fast, on-device recognition with natural language understanding for everyday commands like “Schedule a meeting for tomorrow at 3pm,” “Turn off the living room lights” or “What’s the weather like today?” The underlying models are designed for low-latency, privacy-first processing with compact model sizes, often under 10MB, making them ideal for both embedded devices and browser-based deployment.

Phrase Spotted Commands – A phrase spotting demo shows how Sensory’s technology detects specific command phrases such as “turn on lights,” “play music,” “set alarm,” “volume up” and “navigate home,” even in challenging acoustic conditions. Sensory’s phrase spotted command models are highly configurable, enabling brands to deploy custom vocabularies tailored to their devices, applications and domains while maintaining on-device, noise-robust performance.

“Developers want to feel the end-user experience before they commit to an SDK or roadmap,” said Todd Mozer, CEO & Chairman at Sensory. “These new WebAssembly-powered, browser-based demos make it simple to experience the speed, accuracy and privacy of Sensory’s technologies in just a few clicks—and they clearly demonstrate how our voice AI can be embedded directly into a company’s own website or web app.”

Designed for edge-first, privacy-first products

Sensory’s voice AI is optimized for:

• On-device performance – Models are tuned for ultra-low power operation and small footprints, enabling deployment from bare-metal MCUs to complex OS-based systems.
• Privacy and reliability – Audio data can be processed locally, minimizing dependence on network connectivity and reducing exposure of sensitive user data to cloud infrastructures.
• Custom, branded experiences – Using Sensory VoiceHub, developers can create custom wake words, command sets, and domain-specific language models that reflect their brand, use case and target markets. 
• Web delivery and embeddability – The new demos highlight a browser-based path for deploying Sensory experiences online, showing how voice AI can run via WASM inside existing websites, web apps and digital product experiences.

An easier way to evaluate and deploy Sensory technology

The new demos complement Sensory’s VoiceHub tools and SDKs, providing a frictionless starting point for evaluation. Teams can quickly validate performance, explore different interaction patterns and then move into development with Sensory’s production-ready development kits and integration support—whether targeting embedded devices, hybrid edge-cloud deployments or browser-based experiences.

“From consumer devices and appliances to vehicles and industrial systems, customers are standardizing in on-device voice to improve responsiveness and protect user privacy,” added Mozer. “These demos are the fastest way to understand what Sensory voice AI can do for the next generation of products and web-based digital experiences.”

Availability

The new wake word, Speech-to-Text & NLU and phrase spotted command demos are available today at the Sensory demo experience page: sensory.com/product-demo. For developers ready to move beyond evaluation, Sensory offers SDKs, tools and engineering support to help teams integrate on-device voice AI into production devices and web experiences quickly and confidently.

About Sensory Inc.

Sensory Inc. develops fast, accurate, and private on-device AI technologies, powering over 2 billion devices globally from Amazon, Google, Microsoft, Samsung, and many others. With more than 60 patents, Sensory’s innovations in speech recognition, emergency vehicle detection, voice assistants, biometrics, and natural language understanding span automotive, consumer electronics, wearables, medical and more.

Media Contact:
Amanda Defelice
Marketing, Sensory Inc.

[email protected]