Suno, best known for making AI-generated music, has opened a public beta of Speech, a feature that creates spoken audio and background music together. It is available on Suno’s web and mobile platforms. Users can switch off the music if they want speech on its own, The Verge reports.
People can either describe an idea for Speech to generate or supply a written script. Suno product chief Jack Brody called it an audio model that generates voice and music together as one track. The Decoder reports that Suno tested the feature with a small group for a month before making the beta public.
How does Speech work?
In Suno’s Create tab, users choose Speech and then select Simple or Advanced mode. Simple mode takes a description, such as a pirate captain rallying a crew. Advanced mode accepts a custom script and offers controls for the voice’s gender, speaking style and degree of variation between generations.
Suno says Speech can be used for formats such as poems, meditations and bedtime stories. The Verge reports that generated clips can run for up to about eight minutes. Background music is optional, so the feature can also be used for a voiceover without a soundtrack.
The beta has limits
Suno says the feature is still being improved based on user feedback. Brody warned that accents can drift—for example, a British voice may sound Australian—and that pauses may be overly dramatic. Suno has not said how it trained the model, The Decoder notes.
Speech arrives as AI music generators face copyright disputes. The Decoder reports that major record labels have sued Suno and that a Munich court recently rejected the company’s fair-use argument for using copyrighted data. The sources do not say whether Speech uses the same training data as Suno’s music tools, so its training and copyright details remain unclear.
Why it matters
For people making poems, bedtime stories or voiceovers, Speech offers a way to create narration and an optional soundtrack in one tool. But it is still a beta, and Suno has not disclosed how the model was trained, so its reliability and the source of its training material remain unknown.