It understands the key
A third is not always four semitones. Selekt follows the major or minor scale note by note, so a harmony stays musical as the melody moves.
Vocal Stack Harmony Creator
Make harmonies from a full song or isolated vocal. Selekt extracts and shapes your recorded voice for backing-vocal stacks—not an AI-generated singer. Already have a finished song? Upload the complete recording—no separate vocal stem needed. With a confirmed account, Selekt extracts the lead from your existing song, then builds harmonies, doubles, and backing-vocal stacks you can hear with the music. Selekt keeps the words, phrasing, note lengths, breaths, and emotional timing locked to the original performance.
Full-song uploads use cloud vocal extraction and require a confirmed account. A clean isolated vocal gives the best result; separation may leave artifacts.
Watch this 2:46 guide once and you’ll know where to click to make your first harmony. Choose a clip, try sounds, hear them with the music, and shape a voice if you want. Save and download whenever it sounds right.
Want to try it before uploading? Choose Bravest Decisions or Emotional Ballad in the workspace below. Analyze either demo vocal and explore the harmony sounds and controls—no upload needed.
Press play, then use the player’s full-screen button to see every step clearly.
Prefer YouTube? Open the video in a new tabWhy this is different
A third is not always four semitones. Selekt follows the major or minor scale note by note, so a harmony stays musical as the melody moves.
The source timing is evidence, not a suggestion. The harmony copies carry the singer’s exact words, note lengths, phrase shape, and dynamics.
Thin the verse, add doubles in the pre-chorus, and open the full stack in the hook. Phrase switches make that a deliberate arrangement instead of one effect across the song.
Start wherever the song is
Bring an already-recorded track, an isolated vocal, or a vocal and instrumental pair. The harmony follows your performance; you can preview a short passage, then work on the whole song.
Drop the lead straight into Vocal Stack. Pitch, note, phrase, and preview processing stays in your browser.
Hear the harmonies against the music, then download the finished vocal, instrumental, and in-song mix together.
Choose the full-song input right here. After you sign in with a confirmed account, Selekt extracts the vocal and backing track so you can build and audition harmonies without switching tools.
Take the session with you
Questions producers ask
Yes. Selekt analyzes the notes and phrase timing in one lead vocal, then builds tuned doubles and scale-aware harmony lanes from copies of that recording. It keeps your words and phrasing instead of generating an AI singing voice.
Yes. Performance lock is the central rule: note lengths, words, phrase starts, breaths, and dynamics come from your source. Each lane can add a deliberate offset or human looseness, but it does not invent a new performance that wanders away.
No. Confirm the detected key and choose a named preset. Selekt translates choices such as “third above” into the correct major or minor scale notes. Advanced users can still change every lane and phrase.
Yes. Choose the full-song input here. With a confirmed account, Selekt uses cloud extraction to prepare aligned vocals and instrumental in this studio, then builds harmonies from the vocal. You do not need to use a separate stem separator first. A clean isolated vocal gives the best result; separation may leave artifacts. Isolated-vocal analysis itself stays on your device.
Yes. Upload your finished recording through the full-song option. With a confirmed account, Selekt separates the lead vocal and instrumental, then builds harmonies from that performance. You do not need the original vocal stem or project files. Separation may leave artifacts.
Yes. Selekt creates doubles and scale-aware backing-vocal stacks from the recorded lead, aligned to its words and timing. You can audition them with the song’s instrumental rather than working from a detached vocal clip.
Yes. The harmony studio accepts an isolated vocal up to 10 minutes and 200 MB. You can audition a short section for speed, then apply your chosen sound to the complete vocal before downloading. A finished-song upload uses cloud separation and has a 250 MB file limit. Longer recordings can take more time and browser memory to analyze and render.
Yes. Previews are free. Full-song cloud extraction requires a confirmed account; downloading a full-length production pack costs 0.5 credits.
The production pack includes your finished processed vocal and a processing recipe. If you added an instrumental or started with a full song, it also includes the instrumental and a finished vocal-plus-music mix. For a full-song upload, it includes the isolated original vocal too. The WAV files share the same start point for easy alignment.
No. If you use your own recording, Selekt reshapes copies of that vocal into harmonies; it does not clone a singer, generate an AI singing voice, or compose a new song. Full-song uploads use AI-assisted vocal extraction first. This describes this tool’s output, not other audio or visuals you might add to a finished post.
Only need a quick double? Try the focused Vocal Doubler for Clean Double and Triple Lead previews. To pitch-correct a song vocal, try Hard Tune a Song.