Pinned Tweet
We just raised $25M in new funding led by Future Ventures (@VenturesVc), with @HyperplaneVC and @lakestarvc - bringing Modulate’s total funding to $60 million.
Building the next phase of audio-native AI: more models, more developers, more of the stack that actually understands today:
- Ranked #1 on Hugging Face for deepfake detection and transcription
- 600M+ hours of audio analyzed
- 10M+ hours analyzed every month
A foundational layer of the AI stack is getting more powerful. More soon.
Building the next phase of audio-native AI: more models, more developers, more of the stack that actually understands today:
- Ranked #1 on Hugging Face for deepfake detection and transcription
- 600M+ hours of audio analyzed
- 10M+ hours analyzed every month
A foundational layer of the AI stack is getting more powerful. More soon.
3
3
22
225,536
Post-call fraud review tells you what already happened.
Modulate’s Velma flags deepfakes and social engineering while the caller is still on the line, with every signal tied back to the audio.
Catch it mid-call ⚠️
Modulate’s Velma flags deepfakes and social engineering while the caller is still on the line, with every signal tied back to the audio.
Catch it mid-call ⚠️
1
0
0
7,020
ElevenLabs. Coqui. Qwen3-TTS.
Three different voice generation models. One deepfake detector that caught all of them 👀
That's what a Senior Software Engineer at an enterprise company found when he put Deepfake Detect to the test:
⭐ 4.5/5 - "Fast, Accurate Audio Fake on models multiply, detection has to keep pace. That's the job Modulate Deepfake Detect is built for 💯
Three different voice generation models. One deepfake detector that caught all of them 👀
That's what a Senior Software Engineer at an enterprise company found when he put Deepfake Detect to the test:
⭐ 4.5/5 - "Fast, Accurate Audio Fake on models multiply, detection has to keep pace. That's the job Modulate Deepfake Detect is built for 💯
1
0
0
208
Test it for yourself. 1000 free API credits here: shorturl.at/BVWq8
0
0
0
26
Your STT vendor’s WER was measured on someone else’s audio.
Test it on yours:
1. 50–100 real clips
2. Identical settings across every API
3. Score WER vs human transcripts
4. Price the full feature set
Full framework in comments👇
Test it on yours:
1. 50–100 real clips
2. Identical settings across every API
3. Score WER vs human transcripts
4. Price the full feature set
Full framework in comments👇
1
0
0
1,133
The full framework is in Modulate’s Transcription API Buyer’s Guide: modulate.ai/ebooks/transcr…
0
0
1
21
Modulate retweeted
Been having more 1-1s and this one with Mark Frumkin,VP of CS & Data at @modulate_ai was fascinating so, enjoy!
open.spotify.com/episode/7cnEQr…
open.spotify.com/episode/7cnEQr…
0
1
1
24
AI voice agents shouldn’t be left unsupervised.
They face sarcasm, provocation, edge cases, and callers deliberately trying to push them off-script. And most are trained to keep helping, even if that means hallucinating an answer. 🤐
Modulate’s Velma provides real-time, for AI voice agents. It doesn’t join the conversation. It monitors every side of it, and flags to your team, or directly to your AI, when things start going off the rails.
→ Customize it in minutes with your SOPs, code of conduct and guardrails
→ Every alert traces back to specific comments, sounds or events in the conversation
→ Integrates with your audio channels, CCaaS and ticketing system
The black-box agent experience becomes one you can monitor, understand and control.
They face sarcasm, provocation, edge cases, and callers deliberately trying to push them off-script. And most are trained to keep helping, even if that means hallucinating an answer. 🤐
Modulate’s Velma provides real-time, for AI voice agents. It doesn’t join the conversation. It monitors every side of it, and flags to your team, or directly to your AI, when things start going off the rails.
→ Customize it in minutes with your SOPs, code of conduct and guardrails
→ Every alert traces back to specific comments, sounds or events in the conversation
→ Integrates with your audio channels, CCaaS and ticketing system
The black-box agent experience becomes one you can monitor, understand and control.
0
0
0
3,694
#1 on @huggingface ’s Speech DF Arena 🏅
Modulate Velma ranks first on the independent deepfake detection leaderboard, with a 1.1% average equal error rate. The next-best model sits at 2.1%.
Equal error rate (EER) measures how well a model separates genuine human speech . Lower is better.
Why it matters for fraud teams: a cloned voice on a live call doesn’t wait for post-call review. Velma flags synthetic voices in real time, alongside other voice-native fraud signals:
→ Voice mismatch against known account holders
→ Scripted or rehearsed speech patterns
→ Emotional incongruence, like urgency without stress
Every flag ties back to the original audio, so decisions hold up in audits and disputes.
And it does this with 316M parameters. The two models ranked behind it on the leaderboard run at 1B and 3B.
See how Velma stops voice fraud: hubs.ly/Q04BwQTK0
Modulate Velma ranks first on the independent deepfake detection leaderboard, with a 1.1% average equal error rate. The next-best model sits at 2.1%.
Equal error rate (EER) measures how well a model separates genuine human speech . Lower is better.
Why it matters for fraud teams: a cloned voice on a live call doesn’t wait for post-call review. Velma flags synthetic voices in real time, alongside other voice-native fraud signals:
→ Voice mismatch against known account holders
→ Scripted or rehearsed speech patterns
→ Emotional incongruence, like urgency without stress
Every flag ties back to the original audio, so decisions hold up in audits and disputes.
And it does this with 316M parameters. The two models ranked behind it on the leaderboard run at 1B and 3B.
See how Velma stops voice fraud: hubs.ly/Q04BwQTK0
0
1
2
955
“We don’t speak to AI.”
That’s what an AI assistant heard when it called a pharmacy. Then the call ended.
AI agents are calling. Modulate scores whether a caller is likely human or AI from as little as 2.5s of audio.
Humans → humans. Bots → bots.
That’s what an AI assistant heard when it called a pharmacy. Then the call ended.
AI agents are calling. Modulate scores whether a caller is likely human or AI from as little as 2.5s of audio.
Humans → humans. Bots → bots.
2
0
1
936
Our CTO Carter Huffman breaks down how it works: modulate.ai/blog/ai-agents…
0
0
0
32
Senior Director of Customer Advocacy. Mid-market team. Real workload. 👀
Here's what he had to say about building with Modulate's Transcription AI:
⭐ 4.5/5 "Flexible API and Superb Integration"
"Modulate strikes the right balance of features for our small team, where ality are very important."
Flexible API. Fast setup. Reliable at scale - for teams analyzing customer conversations without the overhead. 💯
Here's what he had to say about building with Modulate's Transcription AI:
⭐ 4.5/5 "Flexible API and Superb Integration"
"Modulate strikes the right balance of features for our small team, where ality are very important."
Flexible API. Fast setup. Reliable at scale - for teams analyzing customer conversations without the overhead. 💯
1
0
1
613
Strat building with Modulate. 1000 free credits on us: shorturl.at/BVWq8
0
0
0
26
44% of new tracks hitting Deezer every day are fully AI-generated.
97% of listeners in Deezer’s blind test couldn’t tell.
Modulate AI Music Detection is built to.
Vocals and instrumentals scored separately, in batch or streaming.
Real music deserves a real shot.
97% of listeners in Deezer’s blind test couldn’t tell.
Modulate AI Music Detection is built to.
Vocals and instrumentals scored separately, in batch or streaming.
Real music deserves a real shot.
1
0
1
41,818
Modulate retweeted
New series: Velma in the Wild.
We take famous scenes, run Velma on the raw audio, and watch what it hears. No script. No transcript. Just the sound.
Episode 2: A Few Good Men, the courtroom scene.
Everyone remembers “You can’t handle the truth!”
But the shout isn’t the real that comes after is.
Watch what Velma flags on the way there.
Words give you the script. Voice gives you the performance.
We take famous scenes, run Velma on the raw audio, and watch what it hears. No script. No transcript. Just the sound.
Episode 2: A Few Good Men, the courtroom scene.
Everyone remembers “You can’t handle the truth!”
But the shout isn’t the real that comes after is.
Watch what Velma flags on the way there.
Words give you the script. Voice gives you the performance.
0
1
2
132
Modulate retweeted
Going live in a few hours 🚨
Your next inbound call might not be a person.
William and I are demoing real-time synthetic voice detection: spot the bot mid-call, then route it where it belongs.
Bots one way. Humans the other.
Join us: linkedin.com/events/7519094…
Your next inbound call might not be a person.
William and I are demoing real-time synthetic voice detection: spot the bot mid-call, then route it where it belongs.
Bots one way. Humans the other.
Join us: linkedin.com/events/7519094…
0
1
1
34
#1 on Hugging Face’s Speech DF Arena deepfake detection leaderboard.
Modulate Velma: 1.1% average equal error rate.
Next best: 2.1%.
Deepfake calls don’t wait for post-call review. Velma flags them live ↓
Modulate Velma: 1.1% average equal error rate.
Next best: 2.1%.
Deepfake calls don’t wait for post-call review. Velma flags them live ↓
1
0
0
431
Voice-native fraud detection for banking, insurance, retail and telecom:
modulate.ai/solutions/comm…
modulate.ai/solutions/comm…
0
0
0
24
Guardrails for AI voice agents, ready on day one.
→ Built on 600M+ hours of real-world conversations
→ Customize in minutes with your SOPs and code of conduct
→ Live alerts to your team, or straight to your AI
Details: hubs.ly/Q04BRZjN0
→ Built on 600M+ hours of real-world conversations
→ Customize in minutes with your SOPs and code of conduct
→ Live alerts to your team, or straight to your AI
Details: hubs.ly/Q04BRZjN0
0
1
0
4,419