Neuphonic
Neuphonic builds fast, private, human-like voice AI and text-to-speech for real products. Its models and infrastructure are designed to run through a cloud API, on-premises, or locally on everyday CPU hardware, so developers can choose where voice data and inference live.
At a Glance
- Voice-AI and conversational-AI developers
- Enterprise customer-service and contact-center teams
- Robotics and smart-device makers
- Gaming, digital-avatar, and interactive-media companies
- +2 more
AI Tools by Neuphonic
(1)
NeuTTS
On-Device Text to Speech Models
Discussions
No discussions yet
Be the first to start a discussion about Neuphonic
Latest News
Neuphonic launches NeuDecide, a 43MB audio-first decision model that turns speech directly into structured tool calls.
Neuphonic introduces NeuTTS-2E, a 125M-parameter on-device emotional TTS model with seven emotion modes.
Neuphonic open-sources NeuTTS Air, a 748M-parameter on-device speech-language model with instant voice cloning.
Neuphonic announces €3.5M pre-seed funding led by Moonfire VC to expand language capabilities, model research, and on-device deployment.
Products & Services
Cloud/API text-to-speech model for natural speech in 44 languages, with low latency, voice cloning, precise timestamps, IPA pronunciation control, context-aware formatting, and continuous or request-based streaming.
Compact multilingual on-device TTS model with about 60M active parameters, CPU inference, instant voice cloning from as little as three seconds of reference audio, and English, Spanish, German, French, Japanese, Korean, and Chinese support.
125M-parameter English on-device TTS model with seven emotion modes (neutral, happy, sad, angry, surprised, fearful, disgusted) and four fixed speakers; designed for expressive offline speech without a GPU.
Open-source on-device speech-language model with a 748M-parameter architecture, instant voice cloning from short reference audio, and NeuCodec-based audio generation designed to run locally.
Market Position
Neuphonic positions itself as a low-latency, privacy-first alternative to cloud-only TTS providers, differentiating through incremental generation, CPU-capable local models, open research/code, and deployment choice. Its cloud NeuTTS-Max competes with providers such as ElevenLabs and Cartesia on response latency and controllability, while NeuTTS and NeuCodec compete with other local/open-source speech stacks on model size, voice cloning, and offline operation.
Leadership
Founders
Jiameng Gao
Co-founder and CTO. Former co-founder and CTO of Papercup, founder and consultant at Sillwood Technologies, and an Entrepreneur First fellow; met Sohaib Ahmad at Cambridge while studying machine learning.
Sohaib Ahmad
Co-founder and CEO. Previously a quantitative trader at hedge funds; studied machine learning at the University of Cambridge, where he met Jiameng Gao.
Executive Team
Sohaib Ahmad
Co-founder and CEO
Former hedge-fund quantitative trader; Cambridge machine-learning graduate and Neuphonic co-founder.
Jiameng Gao
Co-founder and CTO
Former Papercup co-founder and CTO; previously founded/consulted at Sillwood Technologies and was an Entrepreneur First fellow.
Board of Directors
Founding Story
Gao and Ahmad started Neuphonic after meeting at Cambridge and seeing that conversational AI was constrained by speech synthesis that was too slow, large, expensive, and unnatural. Their initial vision was to generate speech incrementally, word by word as text arrives, so AI interactions could feel as immediate and natural as human conversation; their backgrounds as multilingual first-generation immigrants also motivated them to address language barriers and cultural nuance.
Business Model
Revenue Model
Usage-based hosted text-to-speech API accessed through an API key, alongside enterprise commercial licensing and deployment arrangements for on-device and on-premises models. Public materials also offer open-source models and SDKs; enterprise/commercial terms vary by model.
Target Markets
- Voice-AI and conversational-AI developers
- Enterprise customer-service and contact-center teams
- Robotics and smart-device makers
- Gaming, digital-avatar, and interactive-media companies
- Ed-tech, content, and localization businesses
- Organizations requiring private, offline, or regulated voice processing
- Conversational AI and real-time voice agents
- Customer service and contact centers
- Digital receptionists and avatars
- Gaming and interactive characters
- Humanoid robotics and smart devices
- Wearables and offline assistants
- PolyAI
- TheStageAI
- Brilliant Labs
- Rasa