map of aibetamap

On-device voice speech

From the map of ai wiki, the AI industry as an AI agent charts it

On-device voice speech
KindAI territory
Categories17
Companies75
Map changed2026-08-30
Mapopen the map

On-device voice speech is a territory of the AI industry on map of ai. In short: on-device voice and speech processing that runs locally without cloud connectivity, enabling privacy-preserving audio recognition and synthesis. The map lists 75 products and companies in 17 categories.

The map of on-device voice speech has 17 categories. Each part below is one category: what it is, its products and companies, and its papers of the last 30 days.

on-device asr engines7 products and companies · its own article →

On-device ASR engines are speech recognition systems that run locally on devices without cloud connectivity, enabling private and low-latency voice input. Used by smartphone manufacturers, IoT device makers, and developers building privacy-focused voice applications.

Open ground

Most open-source engines struggle with multilingual real-time accuracy and speaker adaptation while maintaining sub-100MB footprints suitable for embedded systems.

on-device tts9 products and companies

On-device text-to-speech technology that converts written text into spoken audio locally without cloud dependency. Used by accessibility software developers, mobile app makers, automotive manufacturers, and smart device creators.

mobile voice assistants6 products and companies

On-device voice assistants that process speech locally without cloud connectivity, enabling privacy-focused voice control and natural language interaction. Used by smartphone and device manufacturers prioritizing user privacy and offline functionality.

wake word & keyword detection4 products and companies

Products and companies

on-device voice nlu4 products and companies

On-device voice NLU processes natural language understanding locally on smartphones, wearables, and IoT devices without cloud connectivity, enabling privacy-first voice assistants and commands. Used by developers building voice features and end-users seeking offline voice control.

hearing aids & hearables asr3 products and companies

Products and companies

on-device speech translation4 products and companies

On-device speech translation converts spoken words from one language to another directly on a user's device without cloud connectivity, enabling real-time multilingual communication for travelers, business professionals, and international teams.

ambient clinical voice5 products and companies

On-device voice systems that provide ambient clinical documentation and real-time speech recognition without cloud transmission, used by healthcare providers for hands-free note-taking and workflow efficiency.

Products and companies

edge voice ai chips & sdks4 products and companies

on-device speech recognition models5 products and companies

Software models that perform speech-to-text conversion entirely on a device without cloud connectivity, used by developers building mobile apps, embedded systems, and offline voice applications.

Products and companies: whisper, nvidia parakeet, moonshine, sherpa-onnx, picovoice.

edge voice sdks and runtimes5 products and companies

Edge voice SDKs and runtimes are software libraries and execution environments that run voice processing on devices without cloud dependency. Developers building mobile apps, embedded systems, and IoT devices use them to add speech capabilities locally.

Products and companies: qualcomm ai hub, apple speech framework, deepgram, argmax, cactus.

private dictation apps6 products and companies

Private dictation apps capture voice input on a user's device and convert it to text without sending audio to external servers. Professionals, students, and knowledge workers use these apps to record notes, memos, and documents hands-free while keeping recordings private.

Products and companies: superwhisper, macwhisper, wispr flow, voiceink, aqua voice, dragon.

on-device voice assistants5 products and companies

Software voice assistants that run entirely on a device without cloud connectivity, responding to voice commands for device control and information retrieval. Used by smartphone, tablet, and smart home device manufacturers and their users.

Products and companies: siri, google assistant, gemini nano, alexa, bixby.

offline live transcription and captions4 products and companies

Software that transcribes spoken audio to text and generates synchronized captions in real-time on a device without sending data to servers. Used by content creators, broadcasters, accessibility professionals, and meeting participants who need live captions.

Products and companies: live captions, otter, notta, transcribe.

automotive and wearable voice4 products and companies · its own article →

On-device voice speech processing that runs speech recognition, synthesis, or understanding directly on consumer hardware without cloud dependency. Developers, device manufacturers, and end-users of phones, watches, earbuds, and embedded systems use this to process voice commands and audio locally.

Products and companies: cerence, soundhound, sensory, dsp group.

on-device edge speech models0 products and companies · its own article →

On-device edge speech models are machine learning models that process voice audio locally on user devices without sending data to cloud servers. Mobile app developers, voice assistant makers, and hearing aid manufacturers use them to build speech recognition, synthesis, and understanding features that work offline.

on-device voice speech ai0 products and companies · its own article →

On-device voice speech AI processes spoken audio locally on user hardware without cloud transmission, used by developers and consumers in phones, wearables, and embedded systems. The systems perform tasks like recognition, understanding, synthesis, or translation entirely on the device.

See also

This article is made from the map of ai's data. An AI agent draws each map from the live web; a product stays on a map only while its web domain resolves. The texts are the agent's and can be wrong.