Unisound · Intelligent Cockpit Agent
Full-Scenario Intelligent Dialogue, Redefining Human-Vehicle Interaction
Built on an integrated edge-cloud large model architecture, lightweight in-vehicle models enable offline interaction, while the cloud-based U2 Large Model provides speech recognition, synthesis, and language understanding. It supports natural offline interaction, multiple languages and dialects, voiceprint recognition, and voice cloning for a smarter in-cabin voice experience.
Industry Pain Points
Deep insight into cockpit scenarioses, solved with AI
Weak Intelligent Cockpit Experience
Most automakers suffer from inadequate native in-vehicle voice functions and poor interactive experiences. Without a full-suite cockpit voice solution, they cannot rapidly integrate various intelligent Agents or meet users’ diverse in-vehicle smart interaction needs, undermining the competitiveness of intelligent cockpits.
Cost-Efficiency vs. R&D Dilemma
Fierce automotive market competition forces automakers to balance cost reduction, efficiency improvement and R&D investment. Traditional in-vehicle voice architectures are inflexible with long deployment cycles and high O&M costs, featuring slow function iterations. They fail to quickly launch cockpit intelligent innovations and adapt to fast-changing market competition.
Lack of Independent Core Capabilities
Traditional voice suppliers adopt black-box delivery. Automakers pay high costs but cannot master underlying technologies in depth. This blocks the accumulation of proprietary technical assets and core in-vehicle voice capabilities, limiting differentiated innovation of vehicle models.
Core Products
A fully self-developed product portfolio covering the entire in-cabin interaction stack.
Agent Ecosystem Services
Supports integration with a wide range of third-party Agent ecosystem services, creating a seamless end-to-end user experience both inside and outside the vehicle.
In-Vehicle Capabilities
End-to-end voice interaction capabilities deployed directly in the vehicle, ensuring a natural language interaction experience even in offline scenarioses.
Cloud Intelligence
Powered by the U2 Large Model, delivering powerful speech recognition, speech synthesis, and natural language understanding capabilities.
Core Advantages
Full-stack in-house technology, deep industry expertise
Edge-Cloud Integrated LLM Architecture
Builds an "edge-cloud integrated LLM" architecture that delivers seamless natural-language interaction even without network connectivity, truly elevating cockpit voice interaction to intelligence.
Precise Colloquial Intent Understanding
Precisely understands intent from colloquial user expressions and triggers the corresponding vehicle actions — even vague phrasing is accurately interpreted.
Full-Link Voice Interaction
From ECNR noise reduction, voiceprint recognition and voice wake-up to on-device LLM and cloud brain — a complete full-link voice interaction stack.
Rich Agents Ecosystem
Connects to diverse third-party Agents services — travel advice, children's companionship, professional knowledge, casual chat and more — building a complete experience loop inside and outside the vehicle.
Multi-Language & Multi-Dialect Coverage
Both speech recognition and synthesis support multiple languages and dialects — Mandarin, English, Cantonese, Sichuan dialect and more — meeting the needs of diverse user groups.
Voice Cloning Technology
A single utterance from the user is enough to faithfully replicate their vocal timbre — delivering a natural, personalized voice without compromise.
Partner Customers
Trusted by leading brands, proving solution value
Geely Ruilan 7
A one‑stop turnkey solution that delivers high‑performance pure‑offline interaction, hundreds of cloud‑based scenario skills and large‑model in‑vehicle applications, enabling diverse intelligent cockpit experiences including knowledge Q&A and travel companionship.
Geely Xingrui / Boyue L
It features ultra‑fast and accurate in‑vehicle voice interaction with snappy response, reliable wake‑up and low false‑wake‑up rate for all‑weather dependable interaction. Leveraging core technologies such as multi‑zone recognition, multi‑layer noise reduction, offline wake‑up and recognition, it adapts to various complex driving scenarioses and delivers smooth and stable voice‑user experience.
Daimler Mercedes-Benz
Built upon ultra‑fast and precise voice interaction technologies, it achieves low‑latency response, stable wake‑up and extremely low false‑wake‑up rate. Combined with capabilities including multi‑zone processing, noise reduction and offline recognition, it fits complex driving environments. It supports recognition for multiple languages and dialects, covering full‑scope in‑vehicle scenarioses such as vehicle control, navigation, phones calls and entertainment. Through regular data iteration and rapid hot‑word optimization, it continuously improves voice recognition accuracy and interaction experience.
IM L6
Boasting ultra‑fast and stable in‑vehicle voice capabilities, it adapts to complex driving scenarioses via noise reduction, multi‑zone processing and offline recognition, and supports multi‑language and multi‑dialect recognition. Sustained data iteration guarantees consistent recognition performance. It supports offline‑online convergence and private deployment. The all‑in‑one model completes semantic understanding and multi‑turn dialogue, and is compatible with large‑model integration to power cockpit applications such as knowledge Q&A, AI content generation and in‑trip companionship.
Lynk & Co Z10
Powered by Unisound’s highly flexible atomic capabilities, it enables seamless offline‑online free interaction and high‑precision voiceprint recognition with deeply optimized offline‑online recognition performance. It has undergone real‑vehicle verification on Lynk & Co Z10, supports in‑car voiceprint enrollment and login, and delivers robust anti‑interference recognition performance.
Dongfeng Mengshi
The end‑cloud integrated voice and large‑model application platform enables joint development covering both end‑side and cloud‑side modules. It supports a wide range of cockpit applications including open‑domain chat, vehicle assistant, trip planning and AI wallpaper. Equipped with a vendor‑oriented management backend, it allows flexible configuration of voice capabilities to ensure smooth interaction experience across end‑side and cloud‑side.
Start Your Smart Cockpit Journey
Our industry experts can provide tailored AI solution consulting to support your digital transformation.
Contact Us