Sensory

What Sensory does

Sensory, Inc., founded in 1994 by Todd Mozer and headquartered in Santa Clara, California, develops on-device AI for voice, sound, and biometrics that runs entirely on edge hardware without cloud dependency. Its product portfolio spans wake word detection (TrulyHandsfree, Smart Wakewords, Personalized, and Secure variants), on-device speech-to-text in 40+ languages with model sizes from 5.3MB to 183MB, phrase-spotted commands, micro language models and custom grammars, environmental sound identification (SoundID), text-independent and text-dependent speaker verification, face verification with liveness detection, and the VoiceHub self-service developer platform. Models are optimized for microcontrollers, DSPs, and NPUs at sub-1mA wake word power and 6-16ms biometric inference latency.

The company sells exclusively through a B2B technology licensing model to OEMs, ODMs, and semiconductor partners, with revenue generated via per-unit royalties and one-time licensing fees tied to device volume and feature scope. Technology is pre-integrated on 60+ chip platforms (Qualcomm, MediaTek, Samsung, NXP, STMicroelectronics, Ceva, Infineon, Ambiq) and deployed in over 3 billion devices globally across automotive, mobile/PC/tablet, TV/home electronics, wearables/hearables, medical devices/healthcare, retail/POS, LLM voice agents, and embodied robotics customers. Named enterprise customers include Samsung, LG, Google, Microsoft, Huawei, Motorola, Nokia, GoPro, Lenovo, Garmin, Honda, Bentley, Volkswagen, Harman, Stryker, Edwards LifeSciences, Fujifilm, Philips, Peloton, Honeywell, Panasonic, and Mastercard. Pricing is quote-based and not publicly disclosed; contracts are multi-year and negotiated.

Sensory operates globally through silicon-vendor partnerships and direct enterprise sales, supported by an event-driven GTM at CES, MWC, Embedded World, CONNECTIONS, and NRF. The company appears to be founder-owned with no disclosed institutional or PE backing since a $400,000 seed in 2001 from Manatt Venture Fund and Norwest; one tuck-in acquisition (vocalize.ai, 2019) signals selective inorganic expansion. Headcount is in the 51-100 range.

Name

Sensory

Legal name

Sensory, Inc.

Website

https://sensory.com

Company type

Private

Founded year

1994

Operating status

Operating

Headcount range

51–100 employees

Short description

Sensory, Inc. is a privately held Santa Clara-based AI company that licenses on-device voice, sound, and biometric recognition technology (wake words, speech-to-text, speaker verification, face verification, sound identification) to OEMs, ODMs, and semiconductor partners across consumer electronics, automotive, healthcare, wearables, and IoT.

Ownership category

akta.pro rank

Product category

Embedded Voice AI Software

SIC

Services-Prepackaged Software (7372)

akta.pro primary industry

On-Device Speech & Audio AI (wake word, ASR, enhancement) (HDAAAJAF)

akta.pro secondary industries

Wake Word & Keyword Spotting (HDAAAFAJ), Speech Enhancement & Noise Suppression (AEC/NR) (HDAAAFAG)

Keywords

• Embedded voice AI
• On-device speech recognition
• Wake word technology
• Speaker verification
• Edge biometrics

Headquarters

HQ city

Santa Clara

HQ country

United States

HQ region

North America

Offices1 record

Markets served

GTM type

B2B

Offering type

Software

Cost components

Technology or R&D, Personnel, Marketing or Sales, Operations, Infrastructure

Revenue model

1. Technology Licensing: Sensory licenses its embedded voice, sound, and biometric AI technology to OEM, ODM, and semiconductor companies. Licensing covers wake words, speech-to-text, speaker verification, face verification, and SoundID. Revenue is generated through one-time licensing fees and per-unit royalties, with pricing varying by volume, device type, and feature set. The company has shipped technology in over 3 billion devices.

Pricing tiers

Go-to-market motion3 records

Distribution channels4 records

Marketing channels8 records

Core offering

Sensory licenses embedded on-device AI technology for voice, sound, and biometrics to OEM, ODM, and semiconductor companies. Its product portfolio includes wake word detection (TrulyHandsfree), speech-to-text, phrase-spotted commands, natural language understanding via Micro Language Models, sound identification (SoundID), and biometric authentication (speaker and face verification). The VoiceHub platform allows developers to create custom wake words and commands without ML expertise. All processing runs entirely on-device with ultra-low power consumption, supporting 40+ languages and deployment across 3+ billion devices.

Product overview

Sensory offers a comprehensive on-device voice and AI platform consisting of multiple integrated products. The core portfolio includes wake word technology (TrulyHandsfree) with variants for standard, smart (context-aware), personalized, and secure (biometric) deployment; speech recognition via Speech-to-Text and Phrase Spotted Commands; natural language understanding through Micro Language Models and Custom Grammars; sound identification via SoundID; and biometric authentication including Text-Independent and Text-Dependent Speaker Verification plus Face Verification. The VoiceHub platform serves as a self-service development portal for creating custom wake words and commands. All products share a common architecture optimized for ultra-low power, on-device processing with no cloud dependency, supporting 40+ languages and integration across major chip platforms. The company positions itself as the edge AI specialist powering billions of devices for major global brands.

Differentiator

Problem solved

Functional benefit

Brands

• VoiceHub: Self-service platform for building custom wake words and word-spotted commands.
• TrulyHandsfree
• SoundID

Products and services

• Sensory Smart Wake Word Context-aware on-device wake word AI with flexible wake phrase placement, adaptive thresholds based on noise levels, and conversational follow-up support without requiring repeat wake words. Designed for OEMs building natural voice-activated consumer devices.
• Sensory Wake Word Ultra-low power, on-device wake word detection with ~15-20KB RAM footprint supporting 40+ languages. Pre-built wake words for fast OEM deployment across microcontrollers, DSPs, NPUs, and application processors.
• Sensory Personalized Wake Word User-defined wake words trained in under 15 seconds (4 repetitions) running fully on-device. Language and accent agnostic with optional voice biometric security for personalized OEM voice experiences.
• Sensory Secure Wake Word Combines wake word detection with built-in speaker verification, ensuring devices respond only when the correct user says the trigger phrase. Phrase-matched security with low-power always-listening architecture.
• Sensory Speech-to-Text Real-time on-device transcription with model sizes ranging from 5.3MB to 183MB, supporting 40+ languages. Runs on mobile operating systems, bare-metal chip deployments, and hybrid cloud configurations with reduced bandwidth.
• Sensory Phrase Spotted Commands Recognizes multiple predefined voice commands simultaneously and triggers immediate actions with ultra-low latency. Noise-robust on-device performance for real-world and far-field use cases.
• Sensory Micro Language Models and Custom Grammars Compact natural language understanding engines for embedded systems interpreting user intent locally without servers. Custom Grammars enable precise recognition of structured commands with task completion rates of 93% vs 55% for competitors.
• Sensory SoundID Detects and classifies 16+ environmental sounds in real time including alarms, sirens, baby cries, glass breaks, coughing, and snoring. Runs fully on-device with multi-stage detection and optional neural revalidation.
• Sensory Text-Independent Speaker Verification Authenticates users based on vocal characteristics from any spoken phrase without requiring a fixed passphrase. Models optimized for CPU at 0.6M-2.5M parameters with 6-16ms inference latency.
• Sensory Text-Dependent Speaker Verification Ties user identity to a specific spoken phrase with two-stage verification combining lightweight first-pass matching and neural revalidation. Provides high-accuracy phrase-matched security on-device.
• Sensory Face Verification Fast on-device facial authentication with millisecond speed, anti-spoofing liveness detection, and works across varied lighting and camera types. Pairs with voice verification for dual-biometric sign-in.
• VoiceHub Self-service online portal for building custom wake words and voice control models without coding. Supports multiple chip architectures, flexible model sizes, and dozens of languages.

Quantifiable outcome

• 93% task completion rate vs 55% for competitor in third-party microwave voice control test
• +4 more outcomes

Named customers64 records

Segments8 records

Ideal customer profiles7 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

Integration3 records

AI capability7 records

Feature11 records

Partnerships

Seven partnerships are on record, tiered major and core.

• STMicroelectronicsmajorTechnology or Integration · 10 March 2026Sensory featured at STMicroelectronics' AI Partner Wall at Embedded World 2026 (Hall 4A/4A-148). Demonstrates how Sensory Speech-to-Text runs on everyday home electronics using ST hardware platforms.
• QualcommmajorTechnology or Integration · 2 March 2026Collaboration to optimize Sensory Micro speech recognition and biometric engine for Qualcomm's Snapdragon Wear Elite wearable chipset. Enables wake word detection and voice commands on next-generation wearables using Qualcomm's Low Power Island architecture and dedicated eNPU. Samsung confirmed Galaxy Watch will use Snapdragon Wear Elite platform.
• NXP SemiconductorsmajorTechnology or Integration · 22 January 2026Sensory's wake word technology integrated into NXP's edge AI portfolio. Partnership expands voice AI capabilities for software-defined vehicles and IoT devices using NXP processors.
• DurincoreTechnology or Integration · 6 January 2026Sensory provides voice biometrics technology for Durin's MagicKey multi-factor authentication system for home entry. The product combines UWB proximity detection, on-device facial recognition, and voice biometrics. Sensory's technology enables secure voice-based user verification in the smart home access market.
• Ceva NeuPro-Nano EcosystemmajorTechnology or Integration · 6 January 2026Sensory's TrulyHandsfree technology integrated into Ceva's NeuPro-Nano NPU ecosystem for ultra-low-power voice activation in battery-constrained consumer electronics including smart home, automotive, and wearable applications.
• CevamajorTechnology or Integration · 5 January 2026Integration of Sensory's TrulyHandsfree voice activation technology into Ceva's NeuPro-Nano neural processing unit (NPU). Enables ultra-low power, always-on voice interfaces for battery-constrained consumer devices including TWS earbuds, smartwatches, smart remotes, and home appliances. All processing occurs on-device ensuring privacy without cloud connectivity.
• MediaTekcoreTechnology or IntegrationCombining MediaTek's generative AI technology with Sensory's on-device voice AI for next-level entertainment and security in vehicles powered by MediaTek Dimensity Auto. Ephrem Chemaly (GM & VP Automotive) leads the partnership.

Scale indicators7 records

Recent moves6 records

Expansion highlights6 records

Direct peers

• Picovoice: Picovoice builds on-device voice AI (wake words, ASR, speech-to-intent) that runs across MCU/CPU/GPU/DSP with similar privacy-first positioning. It is the closest direct substitute to Sensory's TrulyHandsfree and on-device STT for embedded developers.
• Syntiant: Syntiant develops ultra-low-power neural decision processors and edge audio/vision models targeting always-on wake word, sound detection and biometrics on battery-constrained devices. Its ultra-low-power edge AI focus and audio model portfolio overlap directly with Sensory's wake word and SoundID offerings.
• Audio Analytic: Audio Analytic builds on-device AI for environmental sound recognition (alarms, glass break, baby cry) deployed in smart home, security, and IoT. It directly competes with Sensory's SoundID product line in the embedded sound awareness category.
• Pindrop (ID R&D): Pindrop, through its ID R&D acquisition, offers text-independent and text-dependent speaker verification plus facial biometrics for on-device authentication. It is a direct competitor to Sensory's Text-Independent/Text-Dependent Speaker Verification and Face Verification products.
• Retune DSP: Retune DSP provides voice control and keyword spotting algorithms deployed in smart speakers, hearables and IoT. It is a comparable embedded voice AI vendor with similar target customers in low-power consumer devices.
• Paravision: Paravision provides on-device face recognition and liveness AI for access control, mobile, and automotive identity. It competes directly with Sensory's Face Verification module in the on-device face biometrics category.

Broad incumbents

• Cerence: Cerence is a large, established AI provider for automotive in-cabin experiences spanning voice, vision, and biometrics. While it owns the incumbent automotive voice stack, Sensory is increasingly positioning its smaller, on-device engine as a complementary or alternative layer for the same OEMs.
• SoundHound: SoundHound provides a broader voice AI platform (including automotive voice assistants and SoundHound Chat AI) covering more of the full stack. While not exclusively on-device, it competes with Sensory at the OEM voice experience level for automotive and IoT customers.

Emerging players

• Fluent.ai: Fluent.ai offers embedded, language-agnostic speech understanding that runs on small devices offline. It overlaps with Sensory's Micro Language Models and Custom Grammars for compact, multilingual on-device command recognition.
• Ambiq Micro: Ambiq builds ultra-low-power semiconductors and reference AI models (including a voice frontend) for always-on wearables and IoT. It is a partial competitor and ecosystem partner to Sensory, as it sometimes bundles voice capabilities with its low-power MCUs.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat5 records

Key risks6 records

Key highlights6 records

Customer concentration

• Xx.com/trulyhandsfree
• YouTubeyoutube.com/c/SensoryInc
• Facebookfacebook.com/SensoryInc
• LinkedInlinkedin.com/company/sensory-inc-

Compliance2 records

Revenue estimate

Valuation estimate

Number of profiles

Profiles3 records

Funding overview

Funding rounds1 record

Investors2 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A1 record

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Sensory