Make the most of conversational speech with state of the art voice AI
France
Joined April 2024
- Tweets116
- Following12
- Followers537
- Likes80
Pinned Tweet
🎙️ We’re launching the official pyannoteAI Community Discord
A place for developers, researchers, and Voice AI enthusiasts to:
✔ Get support & share knowledge
✔ Showcase projects & learn from each other
✔ Discuss trends shaping the Voice AI ecosystem
discord.gg/vux8UH9QmV
pyannoteAI joins the 2026 #FrenchTech2030 cohort.
80 startups building strategic tech for France, and voice AI is on the list. Good sign for where voice is heading.
Thanks @LaFrenchTech
pyannoteAI is sponsoring VapiCon 2026, Nov 11-12, Fort Mason, San Francisco.
➕ Hervé Bredin is on the “Building Voice Models for Real Conversations” panel, Nov 12, 1:40 PM.
Come find us if overlapping speakers, multi-party calls, or agent evaluation are on your list of production problems. 20% off: PYANNOTEAI20 @Vapi_AI
Last call: Paris Voice AI Meetup, Oct 14.
pyannoteAI x Gladia x Modal on running diarization, transcription, and infrastructure in production. Panel, Q&A, drinks.
luma.com/paris-voice-ai
Already running pyannote in production? On October 8 at 5 PM CEST, we'll show what changes when you switch to Precision-3: the new input parameters, per-frame probabilities in the output, and the migration steps.
Join live or get the recording ➡️ app.livestorm.co/pyannote-ai…
Speaker diarization (pyannoteAI), Transcription (Gladia), Compute (Modal): three layers of one Voice AI stack.
Oct 14, Paris. Panel + live demo of how they fit together.
luma.com/paris-voice-ai
Precision-3 offers a 14.5% diarization error rate and is 3.3 times faster than the leading alternative.
On October 8, Hervé Bredin will show how we measured it and on which datasets. Join us! 🎥 app.livestorm.co/pyannote-ai…
When customized Voice AI, some teams want every overlap flagged, even with a few false alarms. Others would rather have fewer speaker segments they can trust. Precision-3 lets you tune diarization.
We're walking through live on October 8 at 5 PM CEST. 📅app.livestorm.co/pyannote-ai…
Overlapping speakers break your transcripts. The latency budget is gone once you chain models. Costs climb at scale.
If that's familiar, come talk about it in Paris on Oct 14.
pyannoteAI x Gladia x Modal
luma.com/paris-voice-ai
One week until our live session on Precision-3!
October 8, 5 PM CEST. Hervé Bredin and Jyoti Bisht will walk through the benchmark results against Precision-2, the new operating modes, the input parameters, sensitivity controls, and probability outputs.
If you're running diarization in production and want to know exactly what changed before you upgrade, this is the session.
If you no longer believe in the benchmarks published by various Voice AI model providers, this webinar is for you, as we’ll share Precision-3’s results without any hype or false claims.
Register here: app.livestorm.co/pyannote-ai…
Most diarization benchmarks aren't comparable across providers. Precision-3 uses the same 15-dataset methodology as Precision-2: 10.4% lower error, mostly from fixing wrong-speaker attribution. Full results, including where we don't win.
📊 Explore it: pyannote.ai/blog/precision-3…
Voice AI Meetup, Paris, Oct 14. pyannoteAI x Gladia x Modal on what it takes to run diarization, transcription, and infra in production - not a demo.
Panel + Q&A, then drinks.
luma.com/paris-voice-ai
- VAD_sensitivity goes from -0.8 on restaurant audio to +2.0 on audiobooks, so no single value wins everywhere.
- Crosstalk_sensitivity has a cliff: stay between -1 and 0 for general diarization; pushing it higher makes the model flood the output with overlap that isn't there.
We include some settings that you can copy into your pipeline for transcription, voice cloning, dataset prep, noisy multi-speaker audio, and STT reconciliation.
This post also reflects how we want to work at pyannoteAI. Our researchers are committed to writing accessible content that shares their expertise and shows how to use our diarization models in production, beyond benchmark scores.
Full post: pyannote.ai/blog/tune-vad-cr…
Reading about Precision-3's new parameters is one thing. Knowing when to use them is another.
Our DevRel team walks through every feature on real audio, from tuning vadSensitivity to reading the new probability scores.
➡️ youtu.be/gmhNJixkQw0
What counts as good diarization changes by team. Some need every overlapping stretch flagged, even if that means a few false alarms. Others need short pauses broken out precisely. Some would rather work with fewer, more dependable speaker segments than a complete but noisier set.
We'll cover:
- What changed in Precision-3 and why we built it
- The benchmark methodology behind the accuracy and speaker assignment claims
- The sensitivity controls and per-frame probabilities and the speed, balance and accuracy modes.
- How to migrate from Precision-2
➡️ Save your spot: app.livestorm.co/pyannote-ai…