About

Built by a musician who has been on both sides of the demo

Tag-per-Track exists because good demos go unheard. It was created by Lory Ferrand, an artist who sent demos that never got an answer, and a label co-founder who then understood why.

Try it on a track

First analysis free, no card and no account. See a sample report.

Where it comes from

"As an artist, I sent countless demos into the void. As a label co-founder, I realised why: the listening bottleneck."

From the artist's side, silence feels like rejection. From the label's side, it is arithmetic: a small team, a growing inbox, and only so many hours to listen between everything else a label has to do. Nobody decides to ignore a good track. It simply never gets played.

Tag-per-Track is the tool I wanted on both sides of that exchange: something that listens to every submission first, so that human attention goes where it is most likely to matter.

What it does today

You give it a track, and a few seconds later you have:

an A&R score with the reasons behind it, readable from three points of view: discovery, signing and beatmaking;
a check for AI-generated music, with a confidence level;
the artist's live Spotify audience;
the musical profile of the track: tempo, key, genre, mood and instrumentation.

It is available in the browser, through an API, and as an MCP server that lets an AI assistant such as Claude sort a whole folder of demos.

The principles behind it

An unknown artist is never penalised for being unknown. In the Discovery reading, a strong track from an artist under the radar is what the tool looks for, not what it filters out.
Every score comes with its reasons. A number alone does not help anyone decide, so the report says what pushed it up or down.
It helps you decide what to play first, not what to sign. The decision stays with people.
It is honest about uncertainty. When the AI check cannot tell, it says so, and an inconclusive result never lowers a score.
Your music stays yours. Files are deleted once the report is generated and are never used to train AI models.

Who is behind it

Tag-per-Track is designed, built and run by Lory Ferrand, musician and co-founder of the label Black Mountain. It is an independent product, published under the French sole-proprietorship Lory Ferrand EI (SIRET 814 946 208 00034).

Questions, feedback, or a label that would like to try it on its own inbox: contact@tag-per-track.cloud.

See what it says about your track

Drop a file and read its report in a few seconds. The first one is free.

Try it on a track