Where conversations happen: pyannoteAI and Argmax bring speaker diarization to the edge

Life is made of conversations — at home, at work, on the street, in classrooms… everywhere.

At pyannoteAI, our mission is to bring clarity to the chaos of human speech by placing the speaker at the heart of every conversation — revealing the uniqueness behind every voice in every conversation.

Because conversations happen everywhere, speaker intelligence should be available wherever they occur.

Since day one, pyannoteAI has empowered developers around the world to run our technology on their own terms — through a simple API or fully on-premise.

Today, we’re taking things even further by enabling on-device builders to access the most advanced speaker diarization models ever built.

Introducing on-device diarization — delivered by Argmax, powered by pyannoteAI

Together with our friends at Argmax, we’re proud to bring pyannoteAI’s most accurate commercial speaker diarization models to the edge.

Argmax is setting the standard for efficient, privacy-first on-device AI — enabling real-time voice applications to run anywhere, without compromise. This partnership brings together two category leaders to unlock new possibilities for voice intelligence at the edge.

From voice-enabled workflows to (soon) real-time transcription and on-device meeting capture, this collaboration unlocks a wide range of enterprise use cases — where speaker understanding can now happen locally, with low latency and no data ever leaving the device. Whether building field-ready tools, privacy-first communication platforms, or AI agents that operate offline, teams can now integrate state-of-the-art diarization directly into their products — with full control over performance, data, and deployment environments.

Same precision, everywhere

What makes this partnership even more exciting is that we’re not compromising — we’re bringing the exact same precision models that our customers already rely on in the cloud and on-premise, and running them directly on-device. These are the models trusted by top-tier enterprises, built on more than a decade of R&D and real-world feedback.

If you're building voice-first products and want to bring speaker intelligence directly to your users — on-device, at scale, and under your control — get in touch or explore what’s possible with pyannoteAI and Argmax.