Alibaba’s HappyShrimp launches a mobile app; media trial explores songwriting and sharing

Zhidongxi tried mood prompts, original lyrics and reference styles, recording generation times of about one to two minutes while noting limitations in vocal nuance and melody.
Editorial correction (November 16, 2026): Comments on musical style and listening impressions reflect a media trial. Listening alone cannot determine whether a work raises copyright issues or prove that the generative model performed real-time news searches.

Ask AI to write a song in one sentence, and get a finished piece in one minute.
“The chips are still computing, the world is still growing; at the next launch event, though, we will remember that familiar figure.” I asked AI to write a song for Jensen Huang. He did not cry; I did.
Zhidongxi reported on November 13 that Alibaba’s AI music creation platform HappyShrimp officially launched its app that day, becoming available on both iOS and Android.

▲ Screenshot of the HappyShrimp app’s iOS home screen
In mid-August, Alibaba launched its HappyShrimp 1.0 AI music model, initially focused on the web. The newly launched HappyShrimp app targets everyday users creating and discovering AI music. Users describe a mood, story or musical style, and AI handles the lyrics, composition, arrangement and vocals.
Zhidongxi tried the HappyShrimp app immediately after launch and found that it could understand relatively complex natural-language input fairly accurately, match an appropriate musical style, and incorporate knowledge and topical events into its creations. Its lightweight workflow was a highlight: a simple two- or three-step process. Generation was fast, taking about one minute to produce a song roughly three minutes long.
First, I entered a description of a feeling: “My cold has mostly cleared up. After making an active effort, releasing the pain brings relief and happiness.” In about one to two minutes, HappyShrimp used that simple sentence to generate the song below, titled “The Morning My Cold Cleared Up”.
The song sounds relaxed and soothing, and its expanded lyrics stay close to the feeling in the prompt: “Breathing deeply, as light as a cloud; all pain reset to zero, happiness awakening in my veins.” The descriptions feel vivid and tangible. Still, the vocals retain an “AI sound”: stable, but lacking the variations in texture of a real human voice.
Next, I raised the difficulty, asking HappyShrimp to write a song on the theme “Jensen Huang announces his resignation; use this as the starting point for a farewell song to an era.” I requested an emotional progression “from lightness and amazement to reluctance and reflection, ending with the unpredictability of time and eras,” and specified the theme: “One person once changed the world, but the world will ultimately keep moving forward.”
This was intended to test whether HappyShrimp could recognize that “Lao Huang” referred to Jensen Huang, keep the music and lyrics focused on the theme, convey a progression of emotions, and demonstrate agent capabilities such as searching topical news.
As shown below, HappyShrimp generated “A Gentle Farewell to Lao Huang” in about two minutes. The author judged that its lyrics correctly identified the news facts and developed the theme with depth and catchy rhymes. The pop style and bold vocal delivery matched Huang’s eventful career. Overall, though, the melody was fairly conventional and, from an ordinary listener’s perspective, not particularly striking.

What if Xu Fei wrote a song for Huang? I added the prompt “a compositional style similar to ‘The Prose Poems My Father Wrote’.” The app also supports uploading lyrics, so I uploaded lyrics I had written and asked HappyShrimp to generate the song again.
The author felt this version came closer to the requested artistic style and had subtler layers than the first two songs. This is a listening assessment; listening impressions alone cannot establish that a work is free of plagiarism or other rights issues.
HappyShrimp also added an image of Huang seen from behind. The visual suggests a technology leader departing the stage, leaving room for reflection.

Besides creating their own music, users of the HappyShrimp app can also select “Make a similar one”. After discovering a song they like in the immersive music feed—for example, an INFJ song—they can reuse its creative idea with one tap, enter their own MBTI type and quickly generate a personalized version.
HappyShrimp already contains AI-created music shared publicly by a group of users, and a chart of popular songs has been generated from those works.

Overall, our experience suggested that, compared with the PC web version, which is better suited to lengthy, detailed creation and managing works, the HappyShrimp app places more emphasis on capturing inspiration on mobile, lightweight creation, discovering content and interactive sharing.
This article comes from the WeChat public account “Zhidongxi” and was written by “Zhidongxi”.
Related topics
After reading
Join the AINRK community
Read how to join the community and connect with people working in AI.




