Join our daily and weekly newsletters for newest updates and exclusive content to cover the industry. Learn more
New York-based York AI Startup Hume reveals the latest empirical interface interface interface (EVI) model of conversation with AIEVI 3 (pronounced “Evere” three, like the character of the Pokémon), which is specified all from customer support and health coaching association.
Evi 3 lets users create their own voices by talking to the model (it’s voice-to-voice / speech-to-speech), and aims to set a new standard for naturalness, expressiveness, and “empathy” according to hume – that is, how users perceive the model and its abilities to Mirror or adjust its own responses, in terms of tone and word choice.
Designed for businesses, developers, EVI 3 expands past humane models by offering more sophisticated, and more intense understanding.
Individual users can associate with it now Hume’s Live Demo on its website and iOS app, but developer access by proprietary program of Hume’s Hume’s (API) is said to be available in “coming weeks,” as a Blog post from company States.
At this point, developers are able to embed EVI 3 on their own customer service systems, creative projects, or virtual assistants – for a price (see below).
My own use of the demo allows me to create a new, custom synthetic voice in seconds based on the qualities I describe it – a mixture of warm and more masculine. Telling it feels more naturalistic and easy with other AI models and sure that stocks are from legacy leaders Siri and Amazon with Alexa.
Who is whWith developers and businesses need to know about EVI 3
EVI 3 in Hume 3 is built for a range of goods – from customer service and in-app creating content of audiobooks and gaming.
It allows users to specify the precise characteristics of personality, quality vowel, emotional tones, and conversation topics.
This means it can do anything from a hot, kind guide to a quirky, misdemeanor of a french plot of squeal cheese from the kitchen. “
EVI 3 power is in the ability to engage in emotional intelligence directly to voice-based experiences.
Unlike traditional chatbots or voice helps greatly in scripting or interactions based on the text, how people induce, withdrawal, withdrawal, giving up, withdrawing people.
However, a large portion of Hume models currently inadequate – and offered by the rivals of the open open surfer and proprietary, or the powerful treatment of a fuel, such as a CEO Company.
However Heme declares this adds such a competition to the Oct-Text-to-Spress model “Before users of users of users from five seconds to audio.
Hume declares that it is primarily in advance of the behavioral protections and behavior before making this feature available. Today, this cloning ability cannot be used on EVI himself, with a hume highlights the changed voice of adaptation.
Internal benchmarks show users like EVI 3 in OpenII’s click of the OpenII
According to one’s own hume tests with 1,720 users, EVI 3 is preferred to Openi’s GPT-4o In each category checked: naturalness, expressiveness, empatelness, handling transition, em emperation audio ”
It usually targets the Gemini family to Google family and the new open source AI Model Firm Sesame From former oculus co-creator Brendan Irima.
It also boasts lower lency (~ 300 milliseconds), strong multilingual support (English and Spanish, with many languages to come), and effectively unlimited customs. As Hume wrote on its website (see screenshot then below):
Most capabilities include:
- Development generation and express text-in-language in modulation.
- Reconcilingallowing the dynamic invasion of the conversation.
- In-talk to promise to forceSo users can adjust the style of talking in real time.
- API-ready architecture (Arrive soon), so developers can join EVI 3 directly to apps and services.
Access to Price and Development
Hume gives flexible, progress-based EVI, octive tts, and expression measurements.
While the specified EVI 7 conserved was not yet announced (marked as TBA), the pattern suggested to use the use of business discounts.
For reference, EVI 2 has changed at $ 0.072 per minute – 30% lower than the previous, EVI 1 ($ 0.102 / minute).
For creators and developers who work with text-to-spect projects, the Octim PST plan from a free tier (10,000 language characters. Here is the collapse:
- SAVE: 10,000 characters, unlimited typical voices, $ 0 / month
- helper: 30,000 characters (~ 30 minutes), 20 projects, $ 3 / month
- CREATOR: 100,000 characters (~ 100 minutes), 1,000 projects, participate in available overage ($ 0.20 / 1,000 characters), $ 10 / month
- iRO: 500,000 characters (~ 500 minutes), 3,000 projects, $ 0.15 / 1,000 additional, $ 50 / month
- measure: 2,000,000 characters (~ 2,000 minutes), 10,000 projects, $ 0.13 / 1,000 additional, $ 150 / month
- business: 10,000,000 characters (~ 10,000 minutes), 20,000 projects, $ 0.10 / 1,000 $ 900 / month
- business: Custom price and infinite use
For developers who work in real voice interactions or emotional analysis, Hume also offers a salary as you go to the plan with $ 20 in free credits and no committed credits. Customers in Long Volume-CTorprise can opt for a dedicated business plan with Data Licenses, Solutions in place, custom support.
Humane History of Emotive Ai AI models
Built by 2021 by Alan Cowen, a former researcher of Google Defermind, Hume refers to bridge the gap between human emotional and AI interaction.
The company has trained its models in an enthusiastic dataset obtained from hundreds of thousands of participants around the world – the arrest is not only in speech and text, but also face expressions and facial expressions.
“Emotional intelligence includes the ability to control goals and preferences from behavior. That is the core of what interfaces are trying to achieve,” Cowen told VentureBeat. Hume’s mission is to create AI interfaces that are more responsive, man like man, and finally more useful – if helping a customer in an app or take the right mix of drama and laughter.
In the early 2024, the company launches EVI 2, offering 40% lower latency and 30% reduction in EVI 1 price, along with new features of provocation customers and conversations.
February 2025 sees the octave debut, a text writing machine for content creators capable of adjusting emotions in text prompts with text prompts.
With EVI 3 currently available for hand exploration and full access to the API around the corner, Hume hopes to allow developers and creators who can with Voice AI.
