Is Ditto Music worth it for an independent artist releasing AI songs?

Ditto Music: five questions before you pay

1. Is the track fully AI, or did a person make part of it?   -> decides which stores and publishing it can reach
2. Do I hold the rights to the audio AND the lyric?           -> depends on the generator plan and its terms
3. Does the voice imitate a real singer?                      -> if yes, stop here
4. How many releases will I actually finish this year?        -> decides whether a yearly plan pays off
5. Do I want publishing and sync, or only streaming?          -> decides which tier makes sense

Ditto Music is worth it for an independent artist with AI songs under two conditions: you will finish more than one or two releases a year, and a person made a real part of each song. If both hold, a yearly subscription with unlimited releases is a sensible model and Ditto has a written AI policy you can plan around. If the whole track came out of a generator, words, melody and voice alike, Ditto is a weaker fit than it looks, and the reason is on Ditto's own support pages rather than in anyone's forum rumour.

Those pages, as read at the time of writing, say two things that matter for this reader. First, some of the stores Ditto delivers to may not accept releases that are fully AI-generated, so an upload Ditto accepts can still miss some of the stores you picked. Second, Ditto's publishing service will not take songs its system identifies as made entirely with AI tools. Both lines hinge on the same word: fully. This page is about that word, what the Ditto model costs you in practice, and where the human part of an AI song is cheapest to add. Policies in this area change, so treat everything here as a list of things to check on Ditto's current pages, not a verdict.

The human part is cheapest in the words

A distributor releases a finished track. Ditto does not write your song, and no distributor will. When a whole track comes out of a generator, the part a person can most easily and visibly own is the lyric: it is text, it can be drafted, rewritten, dated and kept.

That is where we come in, and only there. The lyrics generator takes a short brief (point of view, emotion, structure, hook, the story in your own words) and returns a title and a full lyric with [Verse] and [Chorus] tags in place. One generation uses 20 credits from a plan's balance, which works out to $0.20 a lyric. It does not sing, it makes no audio, and it has nothing to do with distribution: we are not affiliated with Ditto or any other distributor. The audio happens in a tool you pick, Suno or another generator that sings, under that company's own terms.

Be clear about one thing, because the rest of the article depends on it. A lyric from our generator is also AI output. Pasting it unchanged into a song does not make the song any less "fully AI". What the generator gives you is a fast first draft to rewrite, and the rewriting is the human part. There is no free trial and no single-lyric purchase: the cheapest way in is the $5 monthly plan, which gives the catalogue and the Suno-format prompt generator plus 500 credits every month, 25 drafts. If you plan to release several songs, the $19 lifetime plan is paid once: the catalogue and prompt generator stay yours for good, and 1 500 credits come with it once, which is 75 lyric drafts. Neither is unlimited.

What Ditto says about AI music, as read at the time of writing

Ditto publishes its position in its help centre, in plain language, which is more than some distributors do. Paraphrased, the release article sets out:

• Rights. You must hold the full legal rights to everything in the release. The article names samples, lyrics and AI-generated music specifically. The lyric is on that list, so the rights to your words matter as much as the rights to the audio.
• Voice imitation. No releases that impersonate another artist's voice or persona without permission. A cloned voice is a no, whatever else is fine.
• Spam. No floods of generic releases built to game the algorithms. Fifty near-identical tracks under fifty names is the kind of pattern this targets.
• Store acceptance. Some platforms may not accept fully AI-generated releases, and the article lists which ones. At the time of writing that list included big names such as Apple Music, Audiomack and YouTube Content ID. Check the current list yourself: it is short, it is specific, and it has already changed across the industry more than once.

A separate article on Ditto Music Publishing covers the other half. Collecting societies, it says, do not accept copyright claims for fully AI-generated songs, and if Ditto's system detects that a song was made entirely with AI tools, that song cannot go into its publishing service.

Neither article, at the time of writing, sets a numeric threshold for what counts as "fully". There is no stated percentage, no list of which human contributions qualify. If your plan depends on where that line falls, ask Ditto support directly, in writing, and keep the answer.

For the wider picture of how distributors word these policies and why, see which music distributor accepts AI songs. This page stays on Ditto.

"Fully AI" versus "AI-assisted": the line that decides where a track can go

Put the two Ditto articles together and a practical rule appears. The same distributor can treat two AI tracks very differently:

The table does not promise that track B passes anywhere. Detection systems score audio, and a store can decline a track for reasons nobody explains. What it shows is where your effort moves the needle. For most people making songs with a generator, playing an instrument or singing is a big ask. Rewriting a lyric line by line is not.

Whether that rewriting also makes you the author of the words, in the legal sense, is a separate question, and the answer depends on your country. Copyright offices in some countries, the U.S. Copyright Office among them, have taken the position that text produced from a prompt alone is not protectable, while substantial human reworking may be. How much reworking counts is not something anyone can tell you in advance. Who owns AI-generated music goes through this in more detail. This is not legal advice; where real money or a dispute is at stake, ask a lawyer where you live.

A lyric rewrite, shown line by line

Here is what the human part looks like in practice. A brief in the generator's form:

POV        first
language   en
structure  classic
hook       contrast
vocal      female
emotion    defiance
theme      Night shift nurse on the bus home at 6am, the city waking up
           as she goes to sleep. She refuses to call it a sacrifice.

A draft chorus of the kind a brief like this can come back with. It is competent, and it could belong to anyone:

[Chorus]
The city wakes up as I close my eyes
I'm walking home under morning skies
Don't call it a sacrifice, don't call it pain
I'd choose this road again and again

The same chorus after one pass by the person who knows the story:

[Chorus]
The 6 a.m. bus smells of bleach and bread
You start your day where I go to bed
Don't call it a sacrifice, I'm not that kind
I left the whole ward breathing behind

What changed: of the draft's four lines, only the phrase "Don't call it a sacrifice" survived. The bleach, the bread, the ward and the patients still breathing are not in the brief at all. They came from the person, which is the point. The contrast hook is now a concrete image (your morning, my bedtime) instead of an abstract one, and both rhymes are exact: bread and bed, kind and behind.

Keep the draft, the rewrite and the date together. That file costs nothing and it is the one thing you can show if a store, a collecting society or a distributor ever asks what a person did on this song.

How the Ditto model works, and who it suits

At the time of writing, Ditto sells a yearly subscription in tiers. The tiers differ mainly by how many artist names you can release under and whether publishing and sync are included, and releases are unlimited within a tier. Ditto's pricing page says you keep all of your streaming royalties, and applies a commission to publishing and sync income on the tiers that include them. On top of the plans sit paid extras such as mastering, promotion and ad campaigns, while YouTube Content ID comes with the higher tiers rather than being sold on its own. The prices themselves change, so take the current numbers from Ditto's pricing page rather than from any review, this one included.

What this model means for an AI artist:

• It rewards output. A yearly fee with unlimited releases is cheap per song if you finish ten songs a year and expensive if you finish one. Before you pay, divide the yearly price by the number of releases you will really finish, not the number you hope to.
• It does not reward volume for its own sake. The spam clause cuts against the obvious temptation of a generator plus unlimited uploads. More releases only help if each one is a distinct song someone might choose to hear.
• Publishing is the tier question. If your songs are fully AI, the publishing side of a higher tier may not be usable for them anyway, according to Ditto's own publishing article. Paying for a tier because of a feature your catalogue cannot use is the most avoidable mistake on this page.
• Content ID is a tier feature too. At the time of writing it comes with the higher tiers, and it also sits on Ditto's own list of places that may not accept fully AI releases. Read why YouTube Content ID claims AI tracks before you pay for a higher tier because of it with a generator-made catalogue.

If you are still choosing between models (yearly subscription, pay per release, or a cut of your royalties), the comparison of DistroKid, TuneCore and CD Baby lays those out. The arithmetic is the same whichever company you look at, and the steps around the upload itself are in how to release music on Spotify.

Questions to ask Ditto before you upload

Some answers are on the site; some you will only get from support. Ask in writing and keep the replies with your song files.

1. What happens to my live releases if I stop paying the subscription?
2. How does your system decide a track is "fully AI"? Is there an appeal?
3. If one store declines a release as fully AI, do the other stores keep it?
4. Which stores on my release are currently on your AI exclusion list?
5. For publishing: what would make an AI-assisted song eligible?
6. Which payout methods work for my country, and is there a minimum?

Question 1 matters for every subscription distributor and is easy to skip at signup. Question 3 decides whether a partial refusal is a small problem or a big one. Question 6 is easy to leave until the first payout, and by then it is late.

The paper trail that keeps a release safe

Whatever Ditto decides about a given track, the files below are what you will want if anything is questioned later. None of them costs money.

• The generator's terms at the time you made the audio. Many generators tie commercial use to a paid plan, and the terms change. Save the page, or at least the plan you were on and the date.
• The lyric drafts. The generator output, your rewrites, dated. The example above is the model.
• The voice setting. A note that the vocal was not modelled on a named singer. Ditto's rule against impersonation is one of the few that is not about degree.
• The artwork source. If an image tool made it, the same terms question applies.
• Your answers from support. Especially anything about how a track was classified.

One more point about the lyric. Our generator is one model writing for everyone, so similar briefs can come back with similar lines, and the rights you get in generated text are non-exclusive. We do not check output for uniqueness and we clear nothing against anyone else's rights, as our disclaimer sets out. That is another reason to rewrite: a line only you would write is also a line nobody else is likely to have. This is not legal advice.

The verdict

Take Ditto if you will release several songs a year, a person made a real part of each (the lyric is the easiest part to make yours), the stores you need are not on Ditto's current AI exclusion list, and its payout methods work where you live.

Look elsewhere, or wait, if your plan is one fully generated track, if Apple Music or Content ID is the whole point and your track is fully AI, or if you were counting on the publishing tier for songs a generator made from start to finish.

Either way, the distributor is the last step. The first is a song worth releasing, and a song starts with the words. Write the brief, get a draft (20 credits, $0.20 of a plan's balance), and rewrite it until the chorus has something in it only you would know: open the lyrics generator. The cheapest way in is $5 a month for 25 drafts. If you are going to release more than one song, the $19 lifetime plan on the pricing page gives you 75 drafts once and keeps the catalogue and prompt generator yours for good.