Suno now reads your poem aloud and composes the music behind it in one take

Suno, the app that made its name turning a one-line prompt into a song, now does something stranger with your words: it reads them aloud and writes the music underneath in the same pass. Type a poem, a toast or a bedtime story, describe the voice and the mood you want, and Speech returns one track in which the narration and the score were made together rather than layered afterwards. The hard part of a homemade audio piece has rarely been the voice. It has been making the voice and the music sound as if they belong in the same room.

For anyone who makes wedding toasts, short videos, guided meditations or podcast intros, that removes a step that usually needs two apps and an editing timeline. Today the job means producing narration in one place, finding a royalty-free music bed in another, then nudging volume levels until the music stops drowning the words. Suno’s pitch is that a single model handles both halves, so the soundtrack is written around the speech instead of pasted under it.

Speech has two modes. In Simple mode you describe an idea, such as a pirate captain rallying his crew, and Suno writes the words as well as performing them. Advanced mode takes your own script and adds controls for voice gender, speaking style and how much the delivery varies from take to take. Tracks run to roughly eight minutes, enough for a full bedtime story or a short meditation, and the music is on by default, with a toggle that strips it out for anyone who only wants the narration.

Jack Brody, Suno’s chief product officer, describes it as the first audio model that generates voice and music together as one cohesive track. Synthetic narration on its own is not new. ElevenLabs, Adobe and Google DeepMind all offer text-to-speech that sounds human, but the soundtrack is left to the user, who has to produce it separately and mix it by hand. Suno is entering the field from the opposite side, as a music company adding a voice rather than a voice company adding music.

The rough edges are real, and Suno names some of them itself. The company warns that a British accent can wander off to Australia and back, and that dramatic pauses may be very dramatic. There are no controls for exact duration or word-level timing, and Suno has not said whether the custom voices users can already record for songs work in Speech. It has also not published which languages the feature handles, a gap that matters for listeners outside English-speaking markets, or how many credits a narrated track costs. On the free plan, Suno’s standard terms keep ownership of what users generate and rule out commercial use, so anyone planning a paid podcast or an advert should read the fine print first.

Speech opened to all Suno users on the web and in the iOS and Android apps on October 1, after about a month of testing with a small group. It still carries the beta label, it draws on the same credit plans as song generation, and Suno has given no date for a finished version.

The launch also lands in the middle of a legal fight. Warner Music Group settled with Suno and licensed it in November 2025, but Universal Music Group and Sony Music filed a new suit in September over the v6 music models, accusing the company of copying more than 60,000 recordings and seeking up to $9 billion in statutory damages. That case is pending in federal court in Boston, which means the soundtrack under your bedtime story comes from a company still defending how it learned to make music.

Tags: Generative AI, AI music, text-to-speech, Universal Music Group, Sony Music, Suno