General
Kits AI
- Written by
- Nina Weber
- Published
- Nov 18, 2026
- Read time
- 11 min read
Kits AI is a voice platform made for musicians and audio creators. You record or upload a vocal, pick a voice model, and Kits turns your performance into that voice while keeping your timing, melody and phrasing. Around that core it offers voice cloning from your own recordings, vocal separation, text to speech and a few production helpers such as stem splitting and mastering.
Most voice tools that marketers know start from text. You type a script and get a voiceover back. Kits starts from sound. That makes it a different kind of tool, and for some jobs a better one. If you want a short sung jingle, a demo of an ad song in a different voice, or a clean vocal pulled out of an old recording, Kits was built for exactly that kind of work.
This post covers what Kits AI does, who is behind it, how voice rights work on the platform, where it fits in marketing work, and where a text first audio tool like Content24 AI Audio is the better choice.
What Kits AI does
Kits runs in the browser and is organised around a handful of tools. The names change now and then, but the main parts have stayed the same for a while.
- Voice conversion. Upload a dry vocal and convert it into another voice from the library. The melody and rhythm come from your take, the tone comes from the model. This is the feature Kits is best known for.
- Voice library. A catalogue of ready made voice models. Some are royalty free voices made for the platform, and some are licensed from real artists who agreed to have a model of their voice offered.
- Custom voice models. Train a model on recordings of your own voice, or of a singer who has given you permission, and use it like any other voice in the library.
- Vocal separation and stem splitting. Pull the vocal out of a mixed track, or split a song into parts such as vocals, drums and bass. Handy for remixes and for cleaning up old material.
- Text to speech and voice design. Generate spoken or sung lines from text and shape new voices without recording anything. These features are newer and less central than conversion.
- Production helpers. Reviews also mention voice to instrument conversion and an AI mastering tool. Treat these as extras rather than the reason to sign up.
Who is behind Kits AI
Kits started as a music technology startup and built its name on licensed artist voices. According to Music Business Worldwide, it was backed by Andreessen Horowitz and by musicians such as Steve Aoki and Wyclef Jean. In January 2026 the sample platform Splice bought Kits AI, and the Kits team moved over to Splice. If you already use Splice for samples, expect the two products to grow closer over time.
Kits also says its models are certified by Fairly Trained, a nonprofit that checks whether AI companies trained on data they had permission to use. That certification was one of the reasons Splice gave for the deal. It does not make every output safe to use, but it tells you the company takes training data seriously.
Voice rights on Kits AI: the part you should read first
Voice tools raise one question before any other: whose voice is this, and did they agree? Kits handles this better than many tools, but the rules depend on which kind of voice you use.
- Royalty free library voices. Kits says you get a royalty free license to use the output, as long as you had the rights to whatever you put into the model. In other words, the voice is cleared, but the song or recording you convert still has to be yours.
- Licensed artist voices. These come with the artist's own terms. The Kits artist model license says paying users agree to give the artist a share of gross revenue, set on each artist's page, and that it is the user's job to pay it. Kits also notes that a commercial release in an artist's voice may need approval, and artists can withdraw the license if the use offends them.
- Your own models. If you train a model on recordings you have the rights to, Kits says you can use the output for personal and commercial work. The important words are the rights to. Training on a singer you hired needs their written consent for that specific use.
The simple rule for any brand is this. Only clone or convert into voices you have clear permission to use, and never make a voice that sounds like a real person who has not agreed to it. Imitating a celebrity in an ad is not a grey area. It can break personality rights, advertising law and platform rules all at once, and the damage to trust is worse than any legal letter. Keep the consent in writing, note which model you used, and tell your audience when a voice is synthetic if there is any chance they could think otherwise. Our post on AI compliance covers how to put these rules into a simple team policy.
How to use Kits AI for a marketing project
Here is a typical flow for a short sung jingle or a sonic logo. The same steps work for most conversion jobs.
- Write and record a guide vocal. Someone on the team sings the line as well as they can. Pitch and timing matter more than tone, because the tone will be replaced.
- Keep the take dry. Record without reverb, delay or music underneath. Reviewers agree that clean, isolated vocals convert far better than wet ones. If you only have a mixed track, run vocal separation first.
- Choose a voice. Start with the royalty free library so you avoid revenue shares. Listen to several models with your actual line, not the demo clips.
- Convert and compare. Try two or three voices and small changes in pitch shift. Small changes in the input can make a big difference in the result.
- Fix the weak words. Conversion often struggles with breaths, hard consonants and very high or low notes. Recording one word again is usually faster than fighting the output, so do that and convert just that phrase.
- Mix it properly. Bring the converted vocal into your editor, add the music, and give it the same treatment you would give any vocal. A good mix hides a lot.
- Document the rights. Save a note of which voice model you used, its license terms, and who owns the guide vocal and the music. You will need it when the jingle ends up in a TV spot two years later.
Where Kits AI helps marketing teams
Kits is not a general voiceover tool, but there are a few jobs where it is a very good fit.
- Jingles and sonic logos. Test a short sung hook in several voices before you book a session singer, or use a royalty free voice for small campaigns that would never have a budget for one.
- Song demos for ad agencies. Agencies often pitch a music idea with a rough demo. A converted vocal makes that demo sound closer to the final result.
- Reusing old audio. Separate the vocal from an old brand song to remix it, or clean up a recording where the music is too loud.
- Social content. Short sung versions of a product line, seasonal variations of a known jingle, or a voice that matches a character in a series.
If background music is what you need rather than a voice, a music generator is the better start. We compare options in our guide to AI music for ads and social, and our post on Mureka AI looks at one of them in detail.
Kits AI pricing
Kits has a free plan and several paid tiers that add conversion minutes, download minutes and slots for custom voice models. Third party reviews disagree on the exact prices and on what the free plan includes. Some say the free plan lets you convert but not download, others list a free voice slot. Professional voice cloning sits on the higher tiers. Because the details change and the sources conflict, check the current pricing page on the Kits site before you plan a project around it.
Two practical points. First, the license for artist voices sits on top of your subscription, so a paid plan does not cover the artist's revenue share. Second, some users on review sites report billing trouble after cancelling. Those are individual complaints we cannot verify, but it is a good reason to keep the cancellation email and check your card statement.
Limits of Kits AI
- Quality depends on the input. A poor guide vocal gives a poor result. Kits does not fix pitch problems or weak performances, it only changes the voice.
- Mixed reviews on realism. Some reviewers call the results natural, others hear a robotic edge, especially in long notes and breathy parts. Expect to edit.
- Music first. Spoken voiceovers for explainer videos, product ads or online courses are possible with the text to speech tools, but this is not where Kits is strongest.
- Rights are on you. Kits gives you the license structure. Making sure the input is yours and the revenue share is paid is still your job.
- Not built for teams. There is no shared brand library, approval flow or campaign context. It is a tool for one producer at a time.
Kits AI compared with Content24 AI Audio
Kits and Content24 AI Audio solve different problems, and many teams could use both. Kits is the better tool when you start with a sung performance and want to change the voice, or when you need to separate vocals from a mix. That is music production work, and Kits was built for it.
Content24 AI Audio starts from text. You write or generate a script, pick a voice and a language, and get a voiceover for an ad, a product video or a social post. It sits in the same workspace as the copy, images and video for the campaign, so the script, the visuals and the audio are made from one brief. For dubbing into other languages, our post on dubbing AI explains the workflow, and our guide to ElevenLabs covers one of the voice models behind many text to speech tools.
A simple way to decide: if the job begins with a microphone and a melody, open Kits. If it begins with a script and ends in a campaign, start in Content24.
FAQ
Is Kits AI free?
There is a free plan with limited minutes, so you can test conversion before paying. Downloads, custom voice slots and professional cloning need a paid plan. Check the Kits pricing page for the current limits.
Can I use Kits AI voices in commercial ads?
Royalty free library voices and models trained on your own voice can be used commercially, according to Kits, as long as you own the material you converted. Artist voices come with a revenue share and may need approval. Read the license on each voice page before you publish.
Can I clone any singer's voice with Kits AI?
You should only train models on voices you have permission to use. Kits requires that you have the rights to your training material, and cloning a real person without consent can break the law in many countries. Get written consent that names the use.
Who owns Kits AI?
The music platform Splice bought Kits AI in January 2026, and the Kits team joined Splice. The product still runs under the Kits name.
Is Kits AI good for spoken voiceovers?
It can produce speech, but its strengths are singing and voice conversion. For scripted voiceovers in several languages, a text first audio tool usually gets you there faster.
Conclusion
Kits AI is one of the more careful voice tools around. It was made for music, it licenses its artist voices instead of copying them, and its models carry a Fairly Trained certification. For marketers it is a strong choice for jingles, sonic logos, song demos and vocal cleanup. Use it with clean recordings, stick to voices you have permission to use, and keep the paperwork. When the rest of the campaign needs scripts, voiceovers, visuals and video from one brief, Get Started with Content24.