Suno, the AI music platform best known for turning text prompts into full songs, extended into video on November 17 with Visuals, its first image-and-video feature. Users drag any track from their Suno library into the Visuals tab and generate clips of up to 15 seconds that follow the song's structure, pick up on its lyrics and lip-sync along with them. The web version is live now, with iOS and Android apps to follow in the next few weeks. It is the second expansion beyond music in two months, after the October launch of Speech.
The outputs are aimed at the moment after a song is finished. Visuals can produce lyric videos with custom fonts and colors, album art that matches a track's mood, and short video clips generated from a user's own photos, with a set of original characters available as starting points. Clips can be extended, edited and re-prompted with text on a timeline, and duration and aspect ratio are selectable for different social platforms. The company frames the model as "made for music" — clips react to beat drops and vocal lines rather than playing arbitrary footage over a track.
Under the hood, the model is called H3 Move, and its lineage is notable: Suno built it on top of H3, the open-source video model from Chinese AI startup MiniMax, optimizing it for music-video production and adding its own safeguards. Visual credits are billed separately from song credits and sold in packs that do not expire — according to Suno's pricing page, $14.99 buys 2,500 visual credits, enough for roughly 75 seconds of standard-quality video, while the top $120-a-month subscription includes 20,000 credits, or about 11 minutes. Pro and Premier subscribers get a limited number of trial credits for a limited time.
Suno has also front-loaded the copyright machinery, a sensitive subject for a company that spent two years as the music industry's most prominent defendant. Text and visual filters block prompts and uploads that try to depict specific artists, public figures, copyrighted works and IP, and Suno has said prompts that work on other video platforms may simply not work on Suno by design. Uploads are screened for copyrighted audio through partnerships with Audible Magic, ACRCloud and Musixmatch, generated content is checked against the community guidelines before users see it, and an embedded watermark is added before any Visual can be shared off the platform.
Chief product officer Jack Brody framed the launch around timing rather than novelty: video generation, he argued, had been too slow and expensive for anyone but enterprises until this new generation of models, and Suno's vision had always included it. The competitive logic is visible in the pricing — a separate credit meter for visuals means a second revenue line on top of the music subscriptions, and the music industry's gradual detente with Suno gives it room to sell tools to artists rather than only to their disruptors.
The interesting strategic detail is the supply chain. One of the most commercially successful Western consumer AI music companies is shipping its tuned fork of a Chinese open-weights video model, in the same week its rival platform debates saw Deezer report that one in two listeners now has AI music in their playlists. For independent musicians, Visuals pulls a piece of the music-video budget back into the same subscription they already pay; for the wider generative market, it is another data point that open video models from MiniMax and its peers are becoming the default base layer that consumer products differentiate on top of.
Comments (0)
Log in to join the discussion
Log InNo comments yet