Guides

I checked my catalog against every site it is on

I compared my catalog of 361 tracks with every stock site, library and registry it is on. Here is what was wrong, and how you can check yours.

By Ivan Ryzhko, a composer with 300+ tracks on audio stock and 12 years in production. Updated 21 October 2026. 4 min read.

Why I did it

I have sold production music for 12 years. I started on AudioJungle in 2014. Then Motion Array, and many smaller stock sites. If I count them all, it is more than ten. Then PremiumBeat, and streaming stores through DistroKid. And YouTube Content ID. It is not a store, but it collects money for my tracks too, so it needs the same records.

Today one track is on 6 or 7 places on average. 125 of my tracks are on ten places or more.

For years I kept all of this in a spreadsheet. Then I moved it into tracklog, the catalog app I built. But moving data does not make it right. The data was only as good as my memory. I had a feeling my memory was wrong.

What my catalog knows about each track

• Its status on every platform: live, in review, rejected or not uploaded.
• Its title on every platform, and every title it ever had.
• Tags, description, BPM and other details.
• ISRC, the release it is on, and the release UPC.

This is what made the check possible. To find mistakes, you need one list to compare everything against.

How the check worked

I logged in to each site in a separate browser. Claude, an AI assistant, read each dashboard and compared every item with my catalog. When titles did not match, it compared the audio. When it was not sure, I listened myself. I approved every new title.

The rule was simple: the platform is right, and the catalog follows it.

What I found

My own records were wrong

• Motion Array. My catalog said 110 tracks were live. The site had 181. Ten of them were found only by sound, because they had other titles.
• AudioJungle. 166 in my catalog, 212 on the site. Two tracks I had forgotten completely. One item was renamed back in 2017. This year I sent a new description to the wrong track because of it.
• DistroKid. 7 of 121 "live" marks were wrong. My catalog had ISRC codes for 44 tracks. Now it has 137. One code in my catalog belonged to someone else's recording.
• PremiumBeat. Typos in my own catalog. "Hastle Hard" in my records, "Hustle Hard" on the site. A typo like that is enough to break a royalty match.
• Dates. 359 tracks had a placeholder date. Now 332 have the month and year, taken from the platforms.

Money I was not collecting

• BMI. My tracks are sold under 367 titles that BMI did not have, across 160 works. These are library retitles, renamed items and long stock-site names. When a TV show reports a cue under one of them, the payment may not find me. I sent BMI the full list. One track was never registered at all. Now it is.
• Cue sheets. CNN played one of my tracks 17 times. One show reported a track under its long stock-site name. That is the exact case above.
• Content ID. 661 entries, matched to 222 tracks. One track was there under another name, with 328 claims and $68 earned. I did not know which track it was.

Titles that hurt sales

• 15 tracks on AudioJungle had the same title as another of my tracks. A buyer saw two tracks that looked the same. I gave 14 of them new titles and new tags.
• Envato rejected the first update. With a new title, the file names inside the ZIP must match it. Claude rebuilt the 14 ZIP files and sent them again. All of them were approved in two days.

Ten years of sales

• On AudioJungle, a new track sold about 9 times in its first year in 2015. In 2024–2026 it sells 1 to 1.6 times.
• 37 of my 91 tracks from 2023 on have not sold once.
• My ten best tracks made a third of all my sales.

That changed what I make next. But that is another article.

Who did what

I want to be exact here, because claims about AI are cheap.

• tracklog kept the record: every track, platform, title, old title and code. Everything was checked against it.
• Claude read the dashboards and compared the audio. It wrote the fixes into the catalog, filled in the AudioJungle updates, built the ZIP files and drafted the letter to BMI.
• I logged in, listened to the doubtful pairs, approved every title and sent the letter to BMI.

The AI part is not inside tracklog today. tracklog is the catalog. I am working on connecting the two.

How to check your own catalog without AI

1. Download or copy the list of your items from each site.
2. Put one row per track and one column per site. The free catalog template already has this.
3. Next to each track, write the title each site uses.
4. Check your PRO. Every one of those titles should be registered as an alternate title. See Retitling.
5. Check Content ID. Every asset should match a track you know.

Do it once a year. It took me one day with help. By hand, plan a weekend for every hundred tracks.