


Murf Speech Gen 2 is our most advanced, realistic, and customizable speech model. It challenges the limits of technology, by merging human-like realism with advanced customization capabilities, empowering users to efficiently bridge the gap between their vision and its execution.
Before diving into Gen 2’s capabilities, let’s look at why we built Gen 2, the way we did.
At Murf, we believe creativity is humanity's superpower, deeply personal and uniquely subjective. It is human expertise and subjectivity that transforms a good AI output into exceptional. Hence, while our AI models supercharge your workflow by handling the time-consuming and complex tasks efficiently, they also empower you with customizations to direct the output with a human touch. The last mile towards finished output is loaded with control features, for you to direct our models to work harder and help you achieve your creative vision. This unique approach ensures that everything we build has the creator’s vision at the center, with AI as a catalyst and enabler.
With Gen 2, we want to change the paradigm of how creators evaluate voiceovers; from just ‘Realism’ to how well it bridges the gap between their ‘Vision’ and its ‘Execution’. We believe that the perfect voiceover is not just about sounding real, it is about precisely matching the creator’s vision. Once our AI model creates realistic and authentic voiceovers, the advanced customization features help you harness the full potential of AI, allowing you to mould and modify voiceovers the way you want. Our objective is clear: to ensure that every voiceover delivered by Murf is not just "almost there" but "exactly as intended."
Let’s experience Gen 2’s capabilities.
Realism in AI voiceovers means that AI must understand and reproduce not only the words spoken authentically but also the complex array of human vocal expressions and emotions. Gen 2's neural generative architecture quickly evaluates millions of possible ways to say something, and then narrows them down to the most likely options for you to choose from.
Our second-generation model, designed with Murf’s proprietary, state-of-the-art, generative neural architecture, produces voices that are indistinguishable from human speech. This model has been trained with over 70,000 hours of ethically sourced speech data from diverse demographics and emotion spectrum, resulting in a truly human-like quality in every inflection and rhythm.
Murf Speech Gen 2 operates natively at 44.1kHz sampling rate, allowing it to capture the entire human audible range more precisely. This high-fidelity reproduction ensures that subtle sounds, such as the sibilance in 's' and 'f' sounds, when they occur together, are crystal clear and the voices sound strikingly natural.
To further enhance pronunciation and accent accuracy, we’ve built a deep linguistic modelling layer allowing the speech model to reproduce the subtle nuances of each accent in multiple languages. Rigorous tests by linguists of over 10,000 sentences, revealed a score of over 98.8% word level pronunciation accuracy for our English voice catalogue.
Capabilities Demonstrated: A medley of voices showcasing complex narration styles and words (paralinguistics, compound nouns, brand names).
DocumentaryCapabilities Demonstrated: Thrill, suspense, and intensity with precise pauses and emphatic narration.
Capabilities Demonstrated: Advanced storytelling that captures different forms of expression, such as questions and exclamations.
Capabilities Demonstrated: Speaks loanwords in native or adopted style.
Capabilities Demonstrated: Accurate pronunciation of complex words and abbreviations and precise delivery for longer sentences.
Creators could have completely different takes on how a line is to be spoken. That’s the beauty of creativity and human subjectivity. Gen 2 has been built with a diverse range of voice styles and a host of customization features that give creators the power to direct and shape their voiceover to match their vision.
Gen 2 enables you to choose from a broad spectrum of voice styles - each offering unique pitch, pace, intonation, and emotional depth. This level of customization means that whether you're creating a persuasive business presentation, a captivating audiobook, or an engaging e-learning module, you can choose a style that perfectly aligns with the tone and intent of your content.
While ‘Variability’ allows you to create voiceover versions at a sentence level, creators often need word level subtle nuances to be captured. The Emphasis feature is designed to address this by allowing granular control over the pitch and pace at the word level within any voiceover. Whether you need to underscore the urgency in a safety training module or convey irony in a storytelling session, the Emphasis feature makes it easy to modify vocal elements.
Our ultimate customization feature, 'Say It My Way,' offers the finest degree of control. With ‘Say It My Way’, you can record your rendition of the line to voice-direct the model to capture the intonation, pace, and pitch of your recorded speech. This feature accurately reproduces the exact length and emphasis of each word and pause you make, enabling your selected Murf voice to echo your style.
While ‘Variability’ and ‘Say It My Way’ allows you to create voiceover versions at a sentence level, creators often need word-level subtle nuances to be captured. The ‘Emphasis’ feature is designed to address this by allowing granular control to exaggerate a word. Whether you need to underscore the urgency in a safety training module or convey irony in a storytelling session, the ‘Emphasis’ feature makes it easy to modify vocal elements.
All our advanced customization features - ‘Variability’, ‘Say It My Way’, ‘Emphasis’ - are available for free as part of a limited-time trial, until July 30th, after which users will have to upgrade their plan to continue using them. Enterprise plan users have access to these features by default
Along with Murf Speech Gen 2, we are also announcing the limited launch of Murf Dub. Murf Dub streamlines the creation of multilingual content by quickly delivering high-quality voice dubs that perfectly match the intent of the original, in any language. Whether your audience is in Paris or India, you can create dubbed videos in their language, capturing your message's unique tone and emotion.
Supported by advanced AI and expert linguists, Murf Dub excels at translating complex idioms and contextual nuances. Currently available in five languages - Spanish, Italian, German, French and Hindi - each version undergoes a meticulous review by native speakers to guarantee accuracy and cultural relevance.
At Murf, we recognize the power of AI in transforming voiceover workflows but are equally aware of the ethical responsibilities it brings. As pioneers in AI voice technology, we are committed to upholding the highest standards of ethics to ensure our innovations benefit everyone involved – creatively, responsibly, and fairly.
We ensure our voice technology reflects the diverse world we live in. For instance, our voices span various ages, accents, and cultural backgrounds, making our technology accessible and representative of all users. We do not use public data to train our models, instead relying on proprietary datasets that ensure the highest standards of privacy and integrity. Transparency is key; we openly share how our AI voices are created and used, ensuring users understand the process and trust its integrity. Additionally, we rigorously protect user data, upholding strict privacy standards to secure the personal information entrusted to us.
By prioritizing these ethical principles, Murf not only advances technology but does so with a conscience, enhancing human creativity without compromising values. Read more about our ethical practices here.