Demo management software for record labels

The real bottleneck is listening time

A label does not lack demos, it lacks hours. Submissions pile up in an inbox, a form or a shared folder, and the same few people have to get through them between releases, sessions and meetings. The result is familiar: tracks answered months later or never, and the nagging feeling that something good is sitting unheard at the bottom of the pile.

Tag-per-Track does not replace your ears or your taste. It does the first pass for you, so the listening you do have goes to the tracks most likely to deserve it.

What you get for every demo

• An A&R score from 0 to 100, with the reasons behind it and a plain recommendation: listen first, keep for a second listen, link the artist's profile before deciding.
• A priority level, from "listen first" to "not a priority", so a batch reads as a ranked list rather than a pile.
• An AI check that flags tracks generated with Suno, Udio and similar tools, with a confidence level. How the AI music detector works.
• The artist's live Spotify traction: monthly listeners, followers and popularity, with a career stage (emerging, rising, established).
• The musical profile: BPM, key, mood, genre and instrumentation, plus optional lyrics transcription.

A score that reads the way you work

A scout hunting for unknown talent and a label preparing an offer do not weigh the same things. Each analysis is therefore scored under three reading profiles, and you switch between them instantly:

In Discovery, a finished, catchy track from an artist with fewer than 100,000 monthly listeners is marked as a gem: the kind of submission you want to hear before anyone else does.

From a pile of files to a shortlist

1. Drop the files. One track, or up to 10 at once in batch mode: an EP, an album or a folder of submissions. WAV, MP3, FLAC, AIFF, M4A and AAC are accepted.
2. Read the ranking. For a batch, a comparison view lines the tracks up with their score, tempo and key, and singles out the strongest one as lead single.
3. Keep the history. Every analysis is saved in your dashboard, where you search by title, artist, genre or mood, filter on gems or AI-generated tracks, and sort by score.
4. Share the result. Export a track or the whole history to CSV or JSON, or print a report to PDF for the A&R meeting.

Hand a whole demo folder to your AI assistant

For larger volumes, connect Tag-per-Track to Claude Desktop or Cursor with the MCP server. Point your assistant at a folder and ask in plain language: it analyses every track, checks each artist's traction and returns the demos ranked by priority, with AI suspicions listed apart. It uses the same prepaid credits as the studio, through an API key generated in the API & Agents tab.

Teams with their own tools can call the same analysis through the REST API or the tag-per-track-mcp package.

Where it fits in your workflow

Tag-per-Track is the screening step of your demo pipeline. You keep receiving demos the way you already do, by email, form or shared folder, and run them through Tag-per-Track before the listening session.

It is deliberately not a verdict. The score and the AI check are decision aids: they tell you what to play first, and the decision to sign stays with the people in the room.

Confidentiality of unreleased music

Demos are unreleased material sent in confidence. Files are stored temporarily for the analysis and deleted as soon as the report is generated; only the results are kept in your history. Tag-per-Track does not keep your audio and never uses it to train AI models. Two steps rely on external providers: the AI check, run on a 12-second excerpt, and lyrics transcription, which is optional and sends the full track.

Frequently asked questions

How does Tag-per-Track work with our demo submission form or inbox?

It works alongside it. You keep your current intake (email, form, shared folder) and use Tag-per-Track to analyse and rank what you received, one track at a time, by batch or by whole folder.

How many demos can we analyse at once?

Up to 10 files per batch in the studio. For a whole folder, the MCP server lets an AI assistant such as Claude Desktop triage every track in one request and return a priority-ranked report.

How much does it cost for a label?

There is no subscription. The first analysis is free, then one analysis costs one credit (two with lyrics transcription): €5 for 33 credits or €15 for 100 credits, and credits never expire.

How is the A&R score calculated?

It combines what the audio says about the track (production quality and listening potential) with the artist's live Spotify audience and the loyalty of that audience, weighted differently in the Discovery, Signing and Beatmaker profiles. Every score comes with the reasons behind it. It is a decision aid, not a prediction of commercial success.

What if the artist has no Spotify profile yet?

The track is still analysed and scored on its audio. The report simply indicates that the audience is unknown and suggests linking a profile when one exists.

Can several people on the team use it?

An account is tied to one email address and its credit balance, and its analysis history is synced across devices. Results can be shared with the rest of the team as CSV, JSON or PDF exports.