Liquid AI releases streaming audio codec for the edge
The new Liquid-Codec targets low-latency speech on-device, with multilingual support and a focus on edge deployment.

Liquid AI has published Liquid-Codec, a streaming audio codec aimed at speech applications running on edge hardware. The release is available on Hugging Face and is described as multilingual, positioning it for voice workloads that need to operate close to the user rather than in the cloud.
Audio codecs sit at the foundation of modern speech systems. They compress raw waveforms into compact representations that can be streamed, stored, and — increasingly — fed into generative text-to-speech and speech-understanding models. A codec built for streaming is designed to encode and decode audio incrementally, which keeps latency low enough for real-time conversation.
Why it matters
Much of the recent work in speech AI has focused on server-class models, but a growing share of demand is for voice features that run on phones, wearables, and other constrained devices. A codec optimized for the edge addresses that gap directly:
- Streaming design for responsive, real-time speech
- Multilingual coverage for broader deployment
- Edge-first focus, reducing reliance on cloud round-trips
Liquid AI has built its reputation on efficient, deployment-minded models, and Liquid-Codec fits that pattern as an initial release in a new line. The model is distributed under a custom license, so teams evaluating it for production should review the terms on the model page before building on top of it.
Sources
- Visit
LiquidAI/Liquid-Codec
Hugging Face
More from LiquidAI
All LiquidAI releases →
Liquid AI debuts LFM3-1.7B for edge deployment
A compact 1.7B-parameter multilingual model built to run locally and serve as the base for specialized multimodal variants.

Liquid AI's LFM3-Omni-2B brings any-to-any AI to the edge
The 2-billion-parameter model combines vision, audio and speech-to-speech in a package small enough to run on device.

Liquid AI's LFM2.5 Puts a 24B MoE on Your Device
The new mixture-of-experts model keeps total capacity high while activating only about 2B parameters per token for edge agents.
More in Text → Speech
All Text → Speech →Haru-4B Brings Full-Duplex Speech to Six Languages
A compact 4-billion-parameter model handles real-time, two-way voice conversation across six languages in a single system.

Zyphra releases Zonos3-Mini multilingual TTS
The compact open text-to-speech model supports multiple languages and voice cloning from short audio samples.
OpenBMB's Phonon Brings High-Quality TTS On-Device
The new text-to-speech model promises benchmark-topping voice synthesis that runs locally rather than in the cloud.
0 comments
No comments yet. Be the first to weigh in.