Hi Sean,

We’re thrilled to unveil Empathic Voice Interface 2 (EVI 2), our new foundational voice-to-voice model. It can converse rapidly and fluently with users with sub-second response times, interpret the user’s tone of voice, generate any tone, and even respond to some more niche requests like changing its speaking speed or style.

EVI 2 can maintain a wide range of personalities, including accents and speaking styles. Finally, it can output speech in multiple languages, and will be generally multi-lingual. We’re releasing the EVI 2 API today for all to try, and will continue updating the API as we scale up and improve the model.

Key Features of EVI 2

1. Improved voice quality: More expressive, consistent, natural-sounding speech, with better word emphasis and prosody.
2. Real-time responses: 2x faster with ~500ms - 800ms latency for fluid conversations.
3. Dynamic personalities: Maintains a wide range of engaging, interesting personas.
4. Custom voices: Build on 7 base voices, or adjust attributes like gender and pitch to create your own custom voice.
5. Cost-effectiveness: EVI 2 is 30% cheaper than EVI 1, with a price of $0.0714/min.

Experience EVI 2 Today!

• Try our interactive demo: app.hume.ai
• Access the EVI 2 API: platform.hume.ai
• Watch our Fall Research Update: Meet EVI 2

Introducing the Hume Startup Grant Program

To support the next wave of empathic voice AI innovation, we're launching the Hume Startup Grant Program:

• 🔑 3 months of unlimited EVI API access
• ✨ Technical support access
• 🤝 Co-marketing opportunities!

If you're ready to push the boundaries of voice AI, apply now at hume.ai/grant-program

Explore what's possible in the Hume playground.