A Music Startup Is Betting Against the Prompt Box

A Vinyl Bar in Shibuya is not trying to make another text-to-song slot machine. The more interesting bet is that music apps may get more valuable when they give people knobs, grids, stems, marbles, tempo, and taste instead of one big prompt field.

That is a small consumer app story on the surface. Underneath it is a money story about what investors still think is worth funding in a market drowning in generative AI demos.

Quick Take

• Fact: TechCrunch reports that A Vinyl Bar in Shibuya, a startup founded by former Spotify head of innovation Máuhan M Zonoozy, has raised a $5.5 million pre-seed round.
• Why it matters: This is not a story about one app suddenly beating AI music generators. It is too early for that. The better read is that some consumer investors are still willing to fund interaction, taste, and play as product primitives.

  That matters because a lot of AI music coverage has flattened the question into “can a model generate a song?” The answer is increasing

• Who cares: Consumer app builders should care because this points to a useful design split. If AI makes creation cheaper, the product question shifts from “can we generate?” to “what does the user actually do?”
• Judgment: Fairly hyped as a seed-stage bet that music apps can win through playful human interaction; overhyped if treated as proof that AI music prompting is fading. Fact: TechCrunch reports A Vinyl Bar in Shibuya raised a $5

What happened

TechCrunch reports that A Vinyl Bar in Shibuya, a startup founded by former Spotify head of innovation Máuhan M Zonoozy, has raised a $5.5 million pre-seed round.

The investor list reported by TechCrunch includes Mantis VC, SV Angel, BoxGroup, Quiet Capital, ERA, Common Metal, Gold House, Liquid 2 Ventures, Breakers VC, Systemic Ventures, and former Spotify chief content and advertising business officer Dawn Ostroff, who joined as an advisor.

The company is releasing small music experiments that act more like creative toys than finished studio tools. TechCrunch describes them as experimental “singles” in a broader album of apps and websites.

The examples are concrete:

• Speed Surfer, a Chrome extension that changes speed and filter effects on sound playing in a browser window
• Usersound, a site that turns a username into audio, with letters mapped to notes
• Stacks, a stem player where users create mixes by turning bars on and off in a 3 by 3 grid
• Drops, a marble-and-plaque music toy where falling marbles create notes
• Sampler, a website that turns YouTube videos into music samplers based on beats in parts of songs
• bop, an iOS app for placing musical instruments and effects in a grid to create mixes from a base song, then exporting them to TikTok or Instagram

The company is also reportedly working on a multiplayer version of bop, plus a broader “musical sandbox mixer” that would let users play with different elements of music-making, including instrument creation.

Why it matters

This is not a story about one app suddenly beating AI music generators. It is too early for that. The better read is that some consumer investors are still willing to fund interaction, taste, and play as product primitives.

That matters because a lot of AI music coverage has flattened the question into “can a model generate a song?” The answer is increasingly yes. But music is not only output. It is also identity, fiddling, remixing, sharing, status, learning, and being able to say, “I made that happen.”

A prompt can be powerful, but it can also be a dead end for casual creativity. You type, wait, judge, retry. That is useful for production, but it is not always fun. A grid, a mixer, a marble path, or a stem toggle gives the user a sense of touch. It makes the product feel less like ordering content and more like playing with music.

The money signal is modest but real. A $5.5 million pre-seed round does not prove a category. It does show that investors see room for music apps that are AI-aware without making prompting the whole interface.

Zonoozy’s quoted position in TechCrunch is also worth separating from the hype. He says AI is extraordinarily powerful, but he does not believe the future of music is simply song generation via AI. He argues that if AI makes content more abundant, taste, human perspective, interaction, and participation become more important.

That is a clean thesis. It is also a claim, not a result yet.

Who should care

Consumer app builders should care because this points to a useful design split. If AI makes creation cheaper, the product question shifts from “can we generate?” to “what does the user actually do?”

Music founders should care because the licensing, remix, export, and platform-sharing pieces are likely to matter as much as the interface. The TechCrunch piece describes TikTok and Instagram export for bop, which is exactly where casual music play wants to travel.

AI music companies should care because this is a reminder that generation is not the only lane. There is a difference between making a complete track for a user and giving the user a playful system for shaping sound.

Investors should care because this is a seed-stage test of whether consumer music can feel fresh without leaning on the most obvious AI story. If the apps spread, the value may sit in social play and repeat interaction, not just in model capability.

Creators should care because these tools sit closer to remix toys than replacement engines. That does not remove all rights questions, especially around YouTube sampling and base songs, but it changes the first-order user experience from “make me a song” to “let me mess with sound.”

Bottom line

The useful signal is not “AI music is over.” It is not.

The signal is that prompts are not the only interface for AI-era creativity. As generated content gets easier to make, products that preserve agency, taste, and play may become more interesting, not less.

A Vinyl Bar in Shibuya is early. The funding is small by platform standards. The products sound experimental by design. But the strategy is worth watching because it pushes against the laziest assumption in consumer AI: that every creative app should become a prompt box.

Ride the wave, sure. Just leave room for knobs.

Bandwagon Check

Fairly hyped as a seed-stage bet that music apps can win through playful human interaction; overhyped if treated as proof that AI music prompting is fading. Fact: TechCrunch reports A Vinyl Bar in Shibuya raised a $5

Sources

• A Vinyl Bar in Shibuya is a startup offering fun music apps without any AI prompting
• A Vinyl Bar in Shibuya is a startup offering fun music apps without any AI prompting

By Sean Smith · AI Bandwagon