Songprint

Giving a confidence score with no signup makes it easy to sanity check a track in seconds. As a podcast producer I could see using it on licensed intro music before publishing. Is there any way to see which sections of a file scored as AI, or is it one score for the whole track?

A free detector that returns a confidence score, with 3 checks a day and no signup, is useful as a gut-check before a track goes into a playlist or a client deliverable. The hard part of AI-music detection is short clips and heavily compressed uploads, so I am curious what minimum duration and bitrate you treat as reliable, and whether the score drops when the file is an MP3 re-encode of something that started as WAV. Generator attribution on Pro, naming which model likely made the audio, is the part I would actually pay for, because a bare AI-versus-human label is not enough when someone asks which generator. Batch checks and exportable reports matter more for a catalog than for one-off files.

maybe browser extension can help your audience to check quickly, plus good way to get traffic.

I build audio tools and kept seeing AI-generated tracks passed off as human work in demo submissions and playlists. Existing detectors are either buried behind paywalls or built for labs, not for someone who just needs a quick verdict. Songprint is my answer: paste a link or drop a file, get a confidence score in seconds. The first 3 checks each day are free, no account needed.

I like that the free tier needs no signup and still gives a confidence score rather than just a yes/no verdict. The use case of labels screening demo submissions makes a lot of sense. How does it handle tracks that are partly AI, like a human vocal over an AI-generated backing track? A note on how to interpret mid-range scores would be helpful.

very cool! need to try