Modulate Transcribe just ranked 1 out of 84 models on @huggingface Open ASR Leaderboard.
Beat ElevenLabs. Beat AssemblyAI. Beat Deepgram.
On the industry's MOST trusted transcription benchmark. 🧵
Modulate Transcribe just ranked #1 out of 84 models on @huggingface Open ASR Leaderboard.
Trained on 500M+ hours of real-world, noisy audio - and still costs $0.03/hour.
Incredibly proud of my team at @modulate_ai!
We're outperforming players with 10x our resources, on a
Modulate Transcribe is trained on 500M+ hours of real-world, noisy audio - not clean studio clips.
That's what shows up in production, not just on a leaderboard.
Don't take our word for it.
Test it yourself. 1,000 free credits on us:
How did we do it, while being literally 1000x smaller than many of our competitors? Aside from simply amazing talent, the secret is in the data. @modulate_ai got started on the hardest voice understanding problem out there - real-time understanding of video game voice chats.
@huggingface just listed @modulate_ai as the best transcription model in the world. Best on which metric? Amazingly, nearly all of them.
We're the most accurate model. We're more than 10x *cheaper* than the runner-up models, and our models aren't just good on clear audio, but
Everyone's seen this scene from @netflix Queen Charlotte 👀
Here's what Velma by @modulate_ai sees.
Not just the words - but emotion, sentiment, conversational behaviors, speaker dynamics, and everything else encoded in the voice.
That's the difference between transcription
Voice AI is moving fast. Choosing the right platform shouldn't depend on who has the loudest marketing.
Our latest Voice AI Buyer's Guide covers what actually matters when evaluating solutions:
✅ Accuracy beyond benchmarks
✅ Latency and reliability
✅ Scalability
✅ Security &
Between building voice AI, debugging, and shipping new features, we took a break to catch a bit of the World Cup together! ⚽
Moments like these remind us that great products aren't just built by great engineers - they're built by teams that genuinely enjoy working together.
@modulate_ai has announced the launch of its new AI Music Detection model!
This launch represents a key step in helping the music industry distinguish AI-generated music from human-created content while supporting transparency across the audio ecosystem.
modulate.ai/press-releases…
Most voice pipelines still look like this:
audio → transcribe → text → model → act
The problem is step one. The second you turn audio into text, you throw away tone, hesitation, sarcasm, stress. The signal that told you what the person actually meant.
So everything you
The world's first audio-native AI model is now available as an API.
Velma listens and understands like a human — emotions, tone, intent, rhythm, vocal stress.
Already analyzed 550M+ hours of conversation for Fortune 500s.
Now open to developers. 🧵
the voice intelligence running inside @CallofDuty and @GTAOnline just became available to every developer via API.
same model. same production scale.
@modulate_ai didn't build a demo and work backwards.
its called velma
built by @modulate_ai
What got my attention here is that
GTA Online players are remarkably tolerant of hard gameplay.
What the data suggests they're less tolerant of: abusive voice chat 👀
Players exposed to it show up to 50% higher anger scores - and up to 50% higher odds of logging off within 15 minutes ⏲️
Since implementing
That's a wrap on @CustContactWeek 2026! 🎉
A big thank you to everyone who stopped by our booth, joined a demo, and shared conversations with the team. It was great to connect with so many people shaping the future of customer experience and voice AI.
See you next year!
$300 billion lost to healthcare fraud every year.
Most of it never shows up in the claims data.
It shows up in the phone call before the claim. The telehealth visit with the wrong voice. The social engineering script that bypasses your verification process.
Legacy detection is