Upload & Extend Your Own Audio

Upload an audio file from your own computer and the AI writes a continuation that grows out of the point you choose. The result is built to lead out of your original recording rather than sitting next to it as a separate clip, so a home demo, a loop, or a half-finished idea can be carried forward into a longer piece. Unlike the library-only Extend Track, this workflow starts from your file, so the source does not need to have been made in Music Factory at all.

You can upload in MP3, WAV, OGG, M4A, FLAC, or AAC. The file may be up to 200MB and up to 8 minutes long. It works best with clear, well-recorded audio: a clean take gives the most predictable continuation, whereas a very noisy or distorted file tends to give a looser one. If your recording is quiet or heavily clipped, cleaning it up before uploading tends to pay off in a smoother join.

They do the same job — continue a song — but they start from different places. Extend Track works only on tracks already generated inside Music Factory, selected from your library, while Upload & Extend takes any audio file you upload. Choose based on where your source lives: Aspect · Upload & Extend · Extend Track. Source · Any audio you upload · A generated library track. Made in Music Factory? · Not required · Required. Engines · All versions · All versions. Cost · 20 credits · 20 credits. If your source is already in Music Factory, Extend Track is the more direct route.

Yes. Leave the style, title and prompt fields empty and the run stays a plain continuation of what you uploaded, which is what you want when you simply need a longer version of the same piece. Fill any of them in and the run switches to custom parameters, so you can push the continuation somewhere new — growing a plain loop into a section with its own hook, or moving an instrumental sketch toward a different genre.

You set the point in the uploaded audio where the AI should pick up. Continuing from the end simply extends the piece; picking an earlier point grows a new section out of the middle. As with any extension, clean boundaries give the smoothest result — aim for the end of a phrase, a bar line, or a natural pause rather than mid-word or mid-note, because a tidy edge gives the model the clearest place to take over and keeps the handoff from sounding abrupt.

Upload & Extend works with all Music Factory engine versions, from the classic engines through the latest. The newer engines generally produce cleaner vocals and arrangements, so they are a good default for most uploads. Because your source is an external file rather than a track with a known engine history, there is no single "matching" engine to stay on — pick the newest one for overall quality, or experiment if you want a rougher, more vintage character in the continuation.

Common uses are turning a short voice-memo or guitar loop into a fuller piece, lengthening a track that ended too soon for a video edit, or carrying a co-writer's demo forward to hear where it could go. Anything that starts as a fragment — a hook you hummed, a beat you looped, a verse without a chorus — can be handed to the AI to continue. Once you have a version you like, you can keep extending it or reshape a single part later with Replace Section.

Upload & Extend runs asynchronously, so it is not returned instantly — you upload your file, start the job, and it processes in the background, with the continuation appearing in your library when it is ready. How long that takes depends on load and on the length of your file. You do not need to keep the tab open; the result is saved so you can come back to preview and download it, or extend it again.

The AI-generated continuation you create in Music Factory carries a full commercial license, so you can use the finished Imagera output in videos, ads, podcasts, games, or paid client work. Bear in mind that this only covers the material generated by the tool — if the audio you upload contains someone else's copyrighted recording, extending it does not grant you rights to that original. Credits are consumed when you start each extension.

A little prep goes a long way. Trim any long silence at the start or end so the AI is reading actual music, normalise the level if the file is very quiet, and if the recording is noisy, clean it up before uploading — a clearer file gives a more predictable continuation. If you are extending a loop, make sure it ends on a natural point rather than cutting off mid-phrase, since a clean tail gives the AI the tidiest place to take over and keeps the join from sounding abrupt.

If a continuation wanders, the source audio is usually giving the model little to work from — a very short clip, a heavily processed sound, or a loose performance all leave it less to anchor to. Try uploading a longer or cleaner section so there is more musical context to anchor to, and choose a continue-from point at a clean boundary rather than mid-note. If you deliberately want the piece to stay close to the original, leave the style, title and prompt fields empty; if you filled them in and the result drifted too far, pull those choices back toward what the upload actually sounds like.

Upload & Extend is well suited to turning fragments into finished pieces: a hummed hook becomes a full melody, a guitar loop becomes a longer bed for a video, a co-writer's rough demo gets carried forward to hear where it could go, or a track that ended too soon gets lengthened to fit an edit. Because you start from your own recording, it is also a way to blend a real performance with AI arrangement. Once you have a version you like, keep extending it into new sections, or hand a finished track over to Timestamped Lyrics if it has vocals and you want synced words for a lyric video.