Today, we're releasing Octave: the first LLM built for text-to-speech. Now you can design any voice with a prompt, give acting instructions to control emotion and delivery (sarcasm, whispering, etc.), and produce long-form content on our Creator Studio. Unlike traditional TTS that just “reads” words aloud, Octave understands how meaning affects delivery to generate emotional, human-like speech. Key features: 🎨 Voice Design: Create any AI voice you can imagine with a simple prompt. From "ASMR Southern meditation coach" to "film noir detective" — Octave instantly generates the voices you need for your content. 🎬 Acting Instructions: Octave is the first TTS system that can take natural language instructions to change emotional delivery and speaking style. 🤔 Context-Aware Expression: Trained on 1000x more language than traditional TTS, Octave understands your script like a human actor, delivering realistic emotions, sarcasm, pace, word emphasis, and more. Octave’s first-of-its-kind voice intelligence speaks for itself 😉 In a blind study, Octave outperformed ElevenLabs Voice Design: 🔊71.6% preferred Octave's audio quality 🗣️51.7% found Octave more natural 🎯57.7% said Octave better matched voice descriptions And the best part? Even with superior capabilities, Octave is cheaper than alternatives. Create with Octave today: https://lnkd.in/eQN7hJxT Our blog shares more about Octave, our blind study, and what’s next: https://lnkd.in/eWTxFjBQ #AI #TextToSpeech #VoiceAI #Octave
Hume AI
Research Services
Empathic AI research lab building multimodal AI with emotional intelligence. Experience our API at demo.hume.ai
About us
The foundational voice model with emotional intelligence. Experience our API: https://app.hume.ai
- Website
-
https://app.hume.ai
External link for Hume AI
- Industry
- Research Services
- Company size
- 11-50 employees
- Headquarters
- New York
- Type
- Privately Held
- Founded
- 2021
Locations
-
Primary
New York, US
Employees at Hume AI
Updates
-
How can emotionally intelligent journaling transform your self-reflection? ❤️ 📖 Dive into the insights from Untold’s integration with Hume AI! While many journaling apps focus on helping users document their thoughts, they often fall short when it comes to understanding the emotional nuances behind those words. Untold, an innovative audio journaling app, has changed the game by partnering with Hume AI to integrate their Expression Measurement API. This powerful tool analyzes journal entries in real-time, offering users a detailed emotional readout that fosters moments of self-reflection and growth. With features like "Mood View," powered by Hume’s API, users can track emotional trends and gain deeper personal insights. Untold compared Hume’s API to other LLM prompts and found it outperformed the alternatives, delivering unmatched accuracy and emotional depth. The result? A journaling experience that’s not just about recording thoughts—it’s about truly understanding your emotions. 👉 Ready to explore how empathic AI can enhance your self-care journey? Read more below! https://lnkd.in/eW9-uur2
-
What effect do emotionally intelligent assistants have on consumer shopping experiences? 🤖 🛍️ Read on to hear insights from a new study between The University of Zurich and Hume AI! Traditional AI voice assistants, like Alexa or Google Assistant, excel at functional tasks (e.g., setting reminders or finding information) but often lack the emotional depth needed for meaningful user connections, especially in emotionally driven settings like shopping for experiential products (e.g., scented candles) or customer service interactions. Researchers at the University of Zurich deployed Hume's Empathic Voice Interface (EVI) in an application to test how shoppers would change their consumer behavior while chatting with an emotionally intelligent AI who could talk them through their decisions. The results showed that users strongly preferred an empathic AI over "utility" focused ones for both experiential and functional purchases. Shoutout to University of Zurich researchers Alex Mari, PhD and Ertugrul Uysal, and our very own Jeff Brooks. Read more below! https://lnkd.in/e9iARq6X
-
Did you know that Hume AI's new Octave TTS outperforms ElevenLabs in a blind study with 180 human raters? Octave's outputs were favored over outputs from ElevenLabs Voice Design in terms of audio quality (71.6%), naturalness (51.7%), and how well speech generations matched descriptions of the desired voice (57.7%), across 120 diverse prompts. We're releasing Expressive TTS Arena today for you to compare the two systems against each other and see if you agree! It's a new way to evaluate cutting-edge voice AI systems with natural language instructions and richer text. https://arena.hume.ai/
-
Hear our COO Janet Ho share more about Hume AI's research on speech-language models and what we can look forward to in the near future of voice AI.
“We train foundation models that generate voice and language together, also known as speech language models.” Janet Ho, Hume AI COO, spoke to Forbes' Alexandra York about speech language AI models.
-
Read about how a Fortune 100 automotive company used EVI to explore what drivers want from in-vehicle AI voice companions. 🚗 💨 Hint: drivers want voice AI that’s not just helpful, but also emotionally engaging. Learn more: https://lnkd.in/gRr_GNVy
-
-
You can now transform your expertise into content by having a brief conversation with AI. With Hume’s Empathic Voice Interface, Pressmaster.ai transforms a single AI conversation into weeks of viral content by picking up on what excites users most: ✍️ 93% faster from idea to published content 📈40% increase in content creation activity 🎯 30% increase in demo-to-paid conversions All through AI conversations that make users feel heard and understood, leading to more authentic and impactful content. Learn more: https://lnkd.in/eHnRDiPD
-
-
Introducing OCTAVE. A next-generation speech-language model with emergent capabilities, like on-the-fly voice and personality creation. OCTAVE not only generates high-quality voices but crafts new personalities, accents, and expressions from natural language prompts—all in less than 300ms. What makes OCTAVE different? Our smallest model (3B) can: 🎭 Generate voice, personality, accent, expressions & language from a single prompt 🎯Mimic both voice and personality from audio snippets as brief as 5s 👥Maintain multiple interacting AI speakers in real-time conversation OCTAVE combines frontier LLM capabilities with rich communication, enabling it to follow detailed instructions, use tools, or control an interface. Learn more: https://lnkd.in/dzgVTKi2 We'll be rolling OCTAVE out to trusted partners over the coming weeks. What will you build with OCTAVE?