Intelligent speech / Smart hardware / Industry collaboration
Company overview
Bringing intelligence into the real world.
Built around the SoniCloud speech platform, connecting cloud services with smart devices.
Intelligent speech
Understanding every expression
Smart hardware
Giving products the ability to sense
Software services
Connecting capabilities with work
Manufacturing and delivery
Turning innovation into products
Meet SoniCloud
From a spoken word to everyday work.
Anhui Shengyun Intelligent Technology Co., Ltd. is an innovative technology company built around the SoniCloud speech platform. Our business spans cloud speech recognition, integrated hardware and software solutions, and consumer products.
We provide recognition in Mandarin, English, Southern Min, Cantonese, Japanese, Korean, Russian, and other languages, along with Mandarin and English speech synthesis. Our cloud platform serves partners including Baidu, Huawei, and Honor's mobile assistant. Hardware and software customers include iFLYTEK, Deli, MiMouse, and Yinxiang Biji.
From speech technology to smart hardware for the AI era.
Our technical beginnings
Our predecessor, Xiamen Tiancong Intelligent Software Co., Ltd., was established and introduced a leading embedded speech recognition module in China.
Our technical beginnings2005
Voiceprint engine2008
Smart hardware2016
SoniCloud established2019
National high-tech recognition2021
Industry collaboration2022
Moving into large language models2023
An independent production chain2024
A new way to access AI2025
01 / 09
View the full company timeline
Our technical beginnings
Our predecessor, Xiamen Tiancong Intelligent Software Co., Ltd., was established and introduced a leading embedded speech recognition module in China.
Voiceprint engine
We developed our first voiceprint recognition engine, entering China's voice biometric security market at an early stage.
Smart hardware
Our first wired smart mouse launched. We established a speech lab with Xiamen University, serving companies including Baidu and Huawei, with products reaching hundreds of millions of users worldwide.
SoniCloud established
Anhui Shengyun Intelligent Technology Co., Ltd. was established, providing Baidu Translate with over 40 online speech APIs and handling over 20 million calls per month.
National high-tech recognition
We became a supplier for Huawei's mobile voice assistant and were recognized as a National High-Tech Enterprise that year.
Industry collaboration
iFLYTEK invested in the company, deepening collaboration across speech technology, smart hardware, and commercialization.
Moving into large language models
We developed hardware solutions including Xiaodu scanning pens and shifted toward large language models, launching UMouse smart mouse software and speech transcription software.
An independent production chain
Working with government and industry partners, we established an independently managed production line, secured Baidu AI mouse orders, and joined the supply chains of HP, TPV, and other companies.
A new way to access AI
We completed Series A financing and advanced development of the clawMouse task-oriented AI workspace, using smart hardware as a physical interface to AI and large language models.
Research collaboration
Building on a strong research foundation.
Working with Xiamen University to connect advanced research with real-world applications.
Industry and research
Working with Xiamen University to advance speech technology.
Founded in 2015, XMU Speech Lab brings together researchers from information science, electronic science and technology, and ocean and earth sciences. Its team includes Hong Qingyang, Li Lin, Tong Feng, Xu Binbin, and more than 20 graduate students.
SoniCloud works with the lab on voiceprint recognition, speech recognition and synthesis, speech enhancement, wake words, language identification, and microphone arrays, bringing data and model development, acoustic optimization, and embedded integration to smart devices.
Professor at Xiamen University's School of Informatics and head of the Speech Lab, specializing in speech and voiceprint recognition and contributing to SoniCloud's technology roadmap and research collaboration.
Speech Recognition: Principles and Applications Third edition
From Conformer and Zipformer to speech foundation models, connecting theory with on-device engineering.
Expert perspectives
From speech recognition to natural interaction
Professor Hong Qingyang · Excerpts from public talks
Full talk · From speech recognition to natural interaction6:19
Video notes and sources
Excerpts from public talks, with editorial transitions. The textbook excerpt introduces the second edition of Speech Recognition: Principles and Applications.
SoniCloud's manufacturing facility covers SMT, assembly, testing, and packaging, connecting prototype validation, production coordination, and quality management.
01Reception area at the SoniCloud manufacturing facility02Production floor at the SoniCloud manufacturing facility03SMT and automated production equipment04Product assembly and testing
Our achievements
Innovation that makes its mark.
Core engines, open services, and smart hardware.Every deployment brings speech technology into new areas of industry.
2019Speech APIs in 40+ languages supplied to Baidu TranslateSoniCloud · Technology milestones
2019Monthly speech API calls exceeded 20 millionSoniCloud · Technology milestones
2021Joined Huawei's mobile voice assistant supply chainSoniCloud · Technology milestones
2022Strategic investment from iFLYTEKSoniCloud · Technology milestones
2024Independent smart hardware production line establishedSoniCloud · Technology milestones
2024Secured Baidu AI mouse ordersSoniCloud · Technology milestones
2024Joined the supply chains of HP, TPV, and other companiesSoniCloud · Technology milestones
2025Series A financing completedSoniCloud · Technology milestones
About the original materials
Product screens, certificates, and book covers remain in their original language. Open an item below for an English explanation.
Speech Recognition: Principles and Applications, third edition
The original Chinese book cover credits Hong Qingyang and Li Lin. The English title here is a descriptive translation of the Chinese title, not an assertion that an English edition exists. The cover is reproduced unchanged.
Project: Key technologies and applications for integrated hardware and software in intelligent voice-interaction terminals. Scientific and Technological Achievement Certificate, No. CG044 (2026), issued April 27, 2026, to Anhui SoniCloud Intelligent Technology Co., Ltd. The expert evaluation states that the project reached an internationally advanced technical level. Issuer: China Electronics Energy Saving Technology Association. Joint evaluation organization: International Economic and Technical Cooperation Center, Ministry of Industry and Information Technology.
Certificate HFBD20210463, issued to the company by the Hefei Data Resources Bureau in December 2021, with a stated validity of three years. This is a historical certificate; the image does not establish current validity.
Certificate GR202134000494, issued September 18, 2021, with a stated validity of three years. The approving seals are the Anhui science and technology, finance, and taxation authorities. Reproduced as a historical certificate, without implying renewal.
The plaque names the company as an Anhui Specialized and Innovative Small and Medium-sized Enterprise. Issuer: Anhui Department of Economy and Information Technology. Date printed: December 2025.
Computer Software Copyright Registration Certificate, National Copyright Administration of China. Software: UMouse AI creative writing tool for the large-model mouse, Apple version, V1.0.0. Registration 2025SR0581448; certificate 15237646; April 8, 2025. Rights holder: the company. Rights acquired originally; all rights. Registered following review by the Copyright Protection Center of China under the regulations printed on the certificate. This document is a software copyright registration, not an invention patent.
Computer Software Copyright Registration Certificate, National Copyright Administration of China. Software: UMouse AI spreadsheet management and generation for the large-model mouse, Windows, V1.0.0. Registration 2025SR0447391; certificate 15103589; March 13, 2025. Rights holder: the company. Rights acquired originally; all rights. Reviewed by the Copyright Protection Center of China under the software protection and registration regulations printed on the certificate.
Computer Software Copyright Registration Certificate, National Copyright Administration of China. Software: UMouse data AI spreadsheet processing system for the large-model mouse, Apple version, V1.0.0. Registration 2025SR0447267; certificate 15103465; March 13, 2025. Rights holder: the company. Rights acquired originally; all rights. Reviewed and registered under the software protection and registration regulations printed on the certificate.
Copy of the patent register issued by the China National Intellectual Property Administration. Utility model: A sliding keyboard. Patent ZL202323105669.0; certificate 20978520. Filed November 17, 2023; authorized and announced May 24, 2024. Classification G06F 3/02 (2006.01). Inventor and former holder: Cao Chenxuan. The register records transfer to the company, whose registered address is Room 120, Building B2, Hongfeng Zhigu Entrepreneurship Park, 8 Hongfeng Road, Hefei High-tech Zone, Anhui, postal code 230000, China. The extract states that the patent was valid when issued; it does not establish its current legal status.
The screen includes audio playback, elapsed and total time, playback speed, and Transcript, Summary, Mind map, and Insights tabs. Search transcript keywords finds text within the transcript. Timestamped paragraphs are labeled Speaker 1. The example introduces SoniNote, recording, transcription, translation, noise reduction, meeting minutes, encrypted document sharing, collaboration across displays, tablets, web and phones, secure transmission and storage, and meeting and classroom uses. The sample mentions a one-hour recording transcribed in five minutes and 98% Chinese recognition accuracy; these are statements inside the original sample recording, not new guarantees for this English website.
The film retains its Chinese audio and original imagery, with English subtitles. The recurring on-screen captions describe in-house technology, integrated hardware and software, business digitalization, domestic system compatibility, secure operation, practical applications, and volume partnerships. The software demonstration shows the Chinese AI assistant and the AI PPT action for creating slides. The closing card lists partnership phone numbers 15701171006 and 18075245099 and the address: 9th Floor, Building D9, Phase 1, Zhong'an Chuanggu Science and Technology Park, 900 Wangjiang West Road, Shushan District, Hefei, Anhui, China. The statements reflect the film at the time of recording.
