Music creation used to mean booking a studio, hiring a producer and negotiating with a label for every stem. Today, it can start with a text prompt. Spotify continues to push this direction: the company has expanded its AI remix and covers project in a major partnership with Merlin, the global guardian of independent labels. For independent artists, the move connects the 'vibe coding' trend — describing what you want and letting AI assemble the result — with the everyday reality of streaming playlists.
What is 'vibe coding' and why does it matter for music?
'Vibe coding' is a term popularized by AI researcher Andrej Karpathy in early 2025. In its broadest sense, it means you write a rough idea or a feeling into an AI system, and the system generates the code, image or — in this case — a musical arrangement. You don't have to perfect every line or every hi-hat; you proceed from the vibe. What started as a way for programmers to prototype software is now a creative method in music. Spotify's AI remix feature applies the same logic: you type 'make this track a sleepy lo-fi version' or 'give me a synthwave cover', and the platform generates a new arrangement.
How Spotify expands AI remix and covers project
Spotify has been testing AI-assisted tools since 2024, when it first launched an AI Playlist feature. The current project goes further: users can take any eligible song and create remixes or covers without opening a DAW. The system uses generative models trained on stems and song metadata, then outputs a fully mixed version that fits the original song's mood.
Now, via a partnership with Merlin, this project is expanding beyond the majors. Merlin is a global digital licensing agency for independent labels and distributors. For Spotify, the deal is a way to add a large catalog of independent music to the AI pool. For Merlin's members, it means that their tracks are not locked out of the AI feature. Pending the artist's opt-in, their songs can now be remixed and covered under transparent licensing terms and be tracked on the platform.
What this changes for independent artists
The biggest win is control. In many record-label contracts, AI remixes live in a gray area. With the new partnership, Merlin members get an explicit opt-in/out mechanism, so rights holders decide whether a track can be transformed by AI. Spotify also promises attribution: if a user generates a cover or remix, the original songwriters appear on the release page. The table below compares the old workflow with the new 'vibe coding' workflow.
| Aspect | Traditional remix | AI vibe-coded remix |
|---|---|---|
| Access to materials | Request stems, wait for approval | Work with an existing streaming track |
| Required skills | DAW, mixing, mastering, negotiation | Prompt writing, musical taste, curation |
| Time to first draft | Days or weeks | Minutes |
| Rights handling | Complex sample clearance contracts | Digital attribution and licensing built in |
| Discovery | Released on a single platform | Can be distributed as a new cover/remix on Spotify |
Practical examples: 'vibe coding' a remix in 30 minutes
Let's look at a realistic example. Suppose you are a playlist creator and you want a 'rainy jazz' version of a popular indie song for your late-night mix. With the expanded tool, you don't need Spotify's developer API or original session files. You open the remix interface, select the original track, type 'dark jazz, rain on window, double bass, brushed drums', and the AI returns a finished cover in about two minutes. You can adjust the prompt add '60 BPM' or 'less saxophone' until the vibe matches.
Another example: an independent singer-songwriter wants to launch a cover of a Merlin-affiliated artist. Instead of emailing the label and waiting for a mechanical license, the cover is generated and licensed automatically inside the Spotify ecosystem. The original writer is credited, and the generated cover takes on its own entry while pointing back to the original. That's a very different release cycle from the classic 'remix EP' route. For a label manager, the ability to see which AI covers drive streams to the original track is a useful metric.
A practical checklist for your first AI remix
- Read the opt-in terms of the label or publisher before using a track.
- Start with a clear musical reference: an era, a biome or an emotional color.
- Keep prompts under 10 words; add negative constraints like 'no auto-tune' or 'no heavy drop'.
- Compare the AI result with the original to check whether the harmonic core survives.
- If the AI output is used as a final release, disclose AI involvement in the release notes.
- Never use an artist's actual voice or likeness unless you have a specific agreement.
Copyright, ethics and the limits of AI remixes
This is not a 'robot replaces musicians' story — it's a tool with boundaries. First, a generated cover may not be copyrightable in some jurisdictions because it isn't the result of human authorship. Second, the right of publicity matters: a voice or image of an artist used to generate content can create legal claims. Third, AI models can perpetuate biases: if your prompt asks for something vague, the model might default to Western pop structures and ignore non-Western scales. Practically, this means creators should treat the generated remix as a starting point, not a final release. Record a human piano layer, add your own vocal, edit the arrangement, and then release. Spotify's generated file itself will not create a real 'commercial hit' unless a human curates and polishes the result.
Artists also want transparency about training data. While Merlin's deal sets licensing terms for tracks that are explicitly part of the program, it does not solve every unresolved question about training AI models on all recorded music. For journalists and music professionals, always confirm what data the model used and how rights are credited before advising a client to use these tools.
What to watch next
Every streaming platform is trying to become AI-native. Spotify's move suggests the future is not only 'AI that recommends songs' but 'AI that makes songs you can use in your playlist or your next upload.' For indie labels, the Merlin partnership is an early precedent for how collective rights organizations can negotiate with AI platforms. Expect more details on revenue splits and prompt-based metadata in the coming quarters.
If you are building music-tech products around Spotify data — such as analytics for remix performance or automated cover distributions — API access becomes more important. ASI Biont supports integration with Spotify through its API to help creators connect their web apps to streaming data; learn more at asibiont.com/courses.
Conclusion
Spotify has expanded its AI remix and covers project via a partnership with Merlin, giving independent artists a licensing path and listeners a playful tool. 'Vibe coding' is no longer confined to startups and programmers; it's in your podcast feed, your remix culture and your next late-night playlist. The key is to use it with legal awareness and human taste. Keep an eye on Merlin's and Spotify's official announcements as the feature rolls out to more markets and languages.
Sources: Spotify Newsroom (newsroom.spotify.com), Merlin (merlin.tech), and A. Karpathy's public description of vibe coding (2025).
Comments