Levenza acquires Fluid Inference

We’re thrilled to announce our acquisition of Fluid Inference as part of our mission to elevate every knowledge worker with AI.

Fluid Inference is the applied AI research lab that created FluidAudio and a family of open-source speech models that have been downloaded more than 3.5 million times on Hugging Face. Their models are at the core of apps like FluidVoice, VoiceInk, Spokenly, and Talat — running fast and efficiently enough to work well on everyday devices.

Fluid Inference’s work with NVIDIA made optimized open-source speech models available to developers, while their collaboration with Intel brought speech and language models into AI apps on Intel AI PCs.

Brandon Weng, founder of Fluid Inference, has joined us as Head of Engineering, leading our AI platform and engineering team.

What we’re building together

Working with Levenza should feel as easy and natural as talking to a colleague — which takes models that respond instantly and agents that understand the work well enough to carry it through. Fluid Inference’s expertise bringing speech models into everyday apps complements the agent platform we’ve built for the world's leading professional services and investment firms.

Together, we’ve already accelerated Levenza’s agent development and shipped voice as a new interaction mode. Users can delegate requests on the go or dictate into any app through Levenza, and early voice users are getting through routine admin tasks dramatically faster.

Botless meeting note-taking and transcription are coming next.

Haefa Mansour

Founder & CEO, Levenza