Neuphonic

What Neuphonic does

Neuphonic Limited is a London-based, privately held AI company founded in 2024 that builds ultra-low-latency text-to-speech (TTS) models designed to run entirely on-device on CPU, without GPU or cloud dependency. Its product portfolio comprises three open-source model families — NeuTTS Nano (~120M parameters, sub-200ms latency, 54x real-time factor on CPU), NeuTTS Air (~360M parameters, #1 trending model worldwide on HuggingFace), and NeuCodec (a neural audio codec operating at 0.8 kbps) — all of which support instant voice cloning from as little as 3 seconds of audio and cover 9 languages. The company's core technical differentiator is a documented 10x model size reduction via distillation and pruning that enables deployment on highly constrained hardware such as the Brilliant Labs Halo smart glasses.

The business model is open-core: the models are freely distributed via HuggingFace and GitHub (2M+ and 6K+ cumulative metrics respectively), with paid commercial tiers on the API platform (subscription plus usage-based charges) and OEM/embedded licensing for hardware partners. Go-to-market combines product-led growth through developer self-serve, enterprise field sales targeting privacy-sensitive sectors (healthcare, finance, government) via on-premise deployment, and OEM embedding into devices. Named integrations span AI voice agent platforms (PolyAI, Vapi), real-time communications infrastructure (LiveKit), edge inference SDKs (TheStage AI), consumer hardware (Brilliant Labs Halo smart glasses), and digital human systems (Rapport).

Neuphonic is venture-backed by a single disclosed pre-seed round of £3M (October 2024) led by Moonfire Ventures with Tiny VC, Salica Investments, and Cur8 Capital. The company operates with 1–10 employees and is led by co-founders Jiameng Gao and Sohaib Ahmad. It is incorporated as Neuphonic Limited in England and Wales with its registered office in London.

Name

Neuphonic

Legal name

Neuphonic Limited

Website

https://neuphonic.com

Company type

Private

Founded year

2024

Operating status

Operating

Headcount range

1–10 employees

Short description

Neuphonic is a London-based AI company building open-source, ultra-low-latency text-to-speech models (NeuTTS Nano, NeuTTS Air, NeuCodec) that run on-device on CPU for developers, AI voice agent platforms, and hardware OEMs including smart glasses and wearables.

Ownership category

akta.pro rank

Product category

Text-to-Speech Software

NAICS

Software Publishers (5132)

SIC

Services-Prepackaged Software (7372)

akta.pro primary industry

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

Keywords

• Text-to-speech AI
• Speech language models
• On-device TTS
• Neural audio codec
• Voice synthesis API

Headquarters

HQ city

London

HQ country

United Kingdom

HQ region

Europe

Offices1 record

Markets served

GTM type

B2B

Offering type

Software

Cost components

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

Revenue model

1. API / Platform Access (On-Prem and Cloud): Neuphonic offers its TTS platform via API access and on-premise deployment. The platform has a free tier for non-commercial use and paid tiers for commercial users. Subscription and usage-based pricing is implied based on the Terms of Service referencing paid features, subscription plans, and usage-based charges.
2. Open-Source Models (Free + Potential Enterprise Upsell): Neuphonic's core models (NeuTTS Nano, NeuTTS Air, NeuCodec) are open-source and freely available via HuggingFace and GitHub. Revenue may be derived from enterprise support, customisation services, or premium commercial licensing for large-scale deployments.
3. OEM / Embedded Licensing: Neuphonic integrates its TTS technology into hardware devices (smart glasses, wearables, IoT) through OEM partnerships. Revenue is derived from licensing the technology for embedding in third-party hardware products.

Pricing tiers

Go-to-market motion3 records

Distribution channels5 records

Marketing channels8 records

Core offering

Neuphonic develops and distributes ultra-low-latency text-to-speech AI models (NeuTTS Nano, NeuTTS Air, and NeuCodec) that run in real time on CPU without requiring a GPU or internet connection. The models are offered as open-source downloads on HuggingFace and GitHub, as a hosted TTS API via app.neuphonic.com, and through on-device and on-premise deployment solutions targeted at hardware manufacturers and privacy-sensitive enterprises.

Product overview

Neuphonic is a text-to-speech technology company offering a portfolio of AI voice products. The core portfolio consists of three main models: NeuTTS Nano (smallest, ~120M params, for on-device edge deployment), NeuTTS Air (larger foundation model, ~360M params, #1 trending on HuggingFace), and NeuCodec (open-source neural codec for training speech LMs at 0.8 kbps). These models can be deployed via two solution types: On-Device (for edge/wearable deployment with no cloud dependency) and On-Prem (for enterprise infrastructure with full data sovereignty). All models support instant voice cloning and run on CPU in real-time without requiring a GPU or internet connection.

Differentiator

Problem solved

Functional benefit

Products and services

• NeuTTS Nano Neuphonic's smallest on-device TTS model with approximately 120M parameters, designed to run in real time on CPU without a GPU or internet connection. Supports instant voice cloning from as little as 3 seconds of audio, achieves latency under 200ms with 54x real-time factor, and supports 9 languages for edge and embedded deployment.
• NeuTTS Air A larger CPU-real-time speech foundation model with approximately 360M parameters that delivers higher-quality voice output than NeuTTS Nano while still running on CPU without a GPU. Features instant voice cloning, 24x real-time factor performance, and 9-language support; ranked the #1 trending model worldwide on HuggingFace.
• NeuCodec Open-source neural codec framework that encodes audio at 0.8 kbps, enabling researchers and developers to train high-quality text-to-speech models without building their own codec. Supports upsampling and commercial use.
• On-Device Deployment Solution Deployment package that runs Neuphonic's TTS models directly on user devices without cloud dependency, providing maximum privacy, zero cloud latency, offline capability, and reduced infrastructure cost. Targeted at wearables, IoT devices, mobile apps, automotive, education, gaming, and other privacy-sensitive applications.
• On-Premise Deployment Solution Enterprise deployment package that runs Neuphonic's AI voice systems within an organization's own servers, private cloud, or air-gapped infrastructure. Provides maximum data control, no external dependencies, and customizable infrastructure tailored to enterprise stacks, targeted at healthcare, financial services, and government use cases.
• Neuphonic TTS API Platform Hosted text-to-speech API platform accessible at app.neuphonic.com with an API Playground for testing. Provides commercial-grade TTS with a free tier for non-commercial use and paid subscription tiers for commercial deployments, with usage-based charges per the Terms of Service.

Quantifiable outcome

• NeuTTS Nano achieves 2719 CPU tokens/s with 54x real-time factor (RTF)
• +4 more outcomes

Named customers5 records

Segments4 records

Ideal customer profiles4 records

Technology focussed Yes

API detail

Has API

Yes

API docs

API detail

Core technology

AI maturity

App detail

Integration2 records

AI capability3 records

Feature6 records

Partnerships

Six partnerships are on record, tiered minor and core.

• RapportminorStrategic or Co-development Partner · 1 April 2026Neuphonic and Rapport announced a partnership to develop real-time, photorealistic digital human systems operating on standard CPU hardware. The system combines Neuphonic's low-latency TTS with Rapport's facial and body animation technology, enabling lifelike digital characters for various applications.
• Brilliant LabscoreStrategic or Co-development Partner · 3 March 2026Neuphonic partnered with Brilliant Labs to integrate ultra-low-latency TTS into Halo smart glasses. Neuphonic reduced its model size by 10x via distillation and pruning for deployment on the glasses' constrained hardware. All voice, vision, and sensor data is processed locally on-device, with no data ever leaving the device. The result was the first fully on-device AI smart glasses with real-time voice interaction.
• TheStage AIcoreTechnology or IntegrationNeuphonic partnered with TheStage AI to enable real-time on-device voice across edge devices. Neuphonic provides optimised TTS models for low-latency on-device generation. TheStage AI integrates these models into their inference runtime and developer SDK, enabling manufacturers to deploy high-quality voice capabilities directly on device. Together they deliver a complete stack for fast, private, cloud-independent speech generation.
• PolyAIcoreTechnology or IntegrationNeuphonic works with PolyAI to upgrade the voice layer of their enterprise AI voice agents. Neuphonic's TTS is integrated directly into PolyAI's platform, enabling more natural and expressive speech while meeting the performance demands of real-time, high-throughput customer service deployments at massive scale. Voice interactions feel smoother and more natural, helping turn automated support into human-like conversations.
• VapicoreTechnology or IntegrationNeuphonic partners with Vapi to power interactive AI voice agents, enabling seamless handling of inbound and outbound calls across use cases like customer support, appointment booking, and sales. Neuphonic's TTS is integrated into Vapi-powered agents, enabling fast, natural-sounding speech that works reliably across live phone calls and high-volume deployments.
• LiveKitcoreTechnology or IntegrationNeuphonic partnered with LiveKit to integrate its TTS into the LiveKit real-time voice and video ecosystem. This allows developers to build interactive AI agents capable of real-time audio processing and conversation handling within live WebRTC-based voice environments. The result is a scalable foundation for real-time, voice-driven AI applications.

Scale indicators5 records

Recent moves5 records

Expansion highlights6 records

Direct peers

• ElevenLabs: ElevenLabs is a leading AI voice synthesis company offering high-quality TTS and voice cloning via API. It is the most direct competitor to Neuphonic in commercial voice AI, although it is more cloud-centric while Neuphonic focuses on on-device deployment.
• Resemble AI: Resemble AI provides voice cloning and real-time TTS for enterprise and developer use cases, including on-prem deployment options. It is a comparable commercial voice AI platform to Neuphonic.
• Cartesia: Cartesia builds real-time, on-device speech and audio AI models (including TTS and ASR) with a similar emphasis on low-latency edge inference. It is a direct technical peer to Neuphonic in the on-device speech AI category.
• WellSaid Labs: WellSaid Labs produces AI voice for enterprise content and product teams with a studio-grade API. It targets similar enterprise buyers and competes on quality and reliability.
• Play.ht: Play.ht offers a commercial AI voice generator platform with voice cloning and API access for content creators and enterprises. It directly competes with Neuphonic in the cloud/API TTS market.
• Speechify: Speechify offers consumer and enterprise TTS products with API access. It is a broad TTS peer with overlapping enterprise and developer GTM versus Neuphonic.
• Murf.ai: Murf.ai is an enterprise-focused AI voice generator offering studio-quality TTS across many languages. It targets similar enterprise and content use cases as Neuphonic.

Emerging players

• Coqui (XTTS): Coqui's open-source XTTS model is a free, multilingual voice cloning TTS that directly competes with Neuphonic in the open-source developer ecosystem for self-serve TTS adoption.

Broad incumbents

• Amazon Polly: Amazon Polly is AWS's cloud TTS service, integrated into the broader AWS ecosystem. It represents an incumbent cloud TTS option for enterprises already on AWS, contrasting with Neuphonic's on-device and on-prem focus.
• OpenAI (TTS API): OpenAI offers a cloud-based TTS API as part of a broad AI platform portfolio. It is a broad incumbent with massive distribution and pricing power in the TTS space, though it lacks Neuphonic's on-device capability.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat5 records

Key risks6 records

Key highlights6 records

Customer concentration

• Xx.com/neuphonicspeech
• GitHubgithub.com/neuphonic
• YouTubeyoutube.com/@neuphonicspeech
• LinkedInlinkedin.com/company/neuphonic

Revenue estimate

Valuation estimate

Number of profiles

Profiles2 records

Funding overview

Funding rounds2 records

Investors5 records

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

M&A

Investments

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

Frequently asked questions about Neuphonic