Mubert API - Edit tracks & stems, get consistent music with new engine

Meet the new Mubert API. Edit tracks, swap stems, and generate more consistent music with our latest engine. Create tracks up to 2 hours long, stream music in real time, and integrate Mubert into your pipeline in minutes with Skills.

Add a comment

Replies

Best
Hey all Thanks for checking us out. We worked hard on this release and now you have full control on the music you generate. And you can build with us in few clicks using Claude/Codex Skill. Please share your feedback!

 nice launch.. congrats qq what’s the average response time when requesting real-time stem swaps over the API?

   
Thanks. Straight answer: we didn't optimise for latency in this release — the focus was consistency across a track, which is the thing that makes swapping a stem useful at all.

So a stem swap costs you roughly what generating a track costs. It's not a real-time operation.


Worth separating two things that got announced together: real-time streaming is continuous generation, and stem swapping is an edit on an existing track. The streaming is the low-latency part; the editing isn't, yet.

     Saying it plainly is worth more than the feature, most teams would have let the word streaming cover both and let people find out. It does change what stem editing is for though, because if a swap costs a full generation then it's a revision tool and not something you put behind a user facing control. Worth writing that on the pricing page before someone builds a slider on top of it and eats the bill.

Stems are the part that changes the workflow, everything else is a nicer reroll. The moment I can mute one layer and re-render only that, music stops being a slot machine and becomes an editable asset. For API use the thing that decides it is whether the same seed and prompt returns the same track next month, because branded content gets revised weeks after it ships. Consistency across sessions matters more than consistency inside one.

Hi Product Hunt!

For a long time, music APIs were mostly about sending a request and getting a finished audio file back. But teams building music into real products need much more: track editing, license and user management, usage controls, long-form generation, real-time streaming, and an integration process that does not slow development down.


That is what we focused on with this release. The updated Mubert Music API now supports track editing, individual stems, more cohesive compositions powered by our latest engine, tracks up to two hours long, real-time streaming, and faster integration through Skills for Claude and Codex.


I am especially excited to see what developers build with it, from creator tools and games to wellness, fitness, streaming, and entirely new forms of personalized media.


We would love to hear your thoughts: what product would you add generative music to, and what would you need from an API to make it work?

Hey all,

Thanks for checking us out. This release is about control over what you generate.

You can edit a track after it's made, swap a single stem instead of starting over, and the new engine keeps it consistent across the track.

And you can build with us in a few clicks using the Claude/Codex Skill, so no wading through docs to get started.


Would really love your feedback — especially from anyone who's tried building with a music API before and hit a wall.

the adaptive playlists feel surprisingly natural for AI-made music, no robotic loops here. curious how it handles licensing for client use though.

Licensing is where I got burned before. The licence covered my own content but excluded ads and anything shipped inside a paid app, so I ended up self-hosting a model instead. Does worldwide copyright-protected cover a track my software generates and posts for a paying customer, or only my own use?