SongExtender — Guide to AI Music Creation

Getting Started

Start with a song idea or audio you own or have permission to use. Simple mode takes a short description; Custom mode lets you enter lyrics and a style direction.

Sign in to save your work. Check the selected model, credits and result visibility before choosing Generate. Each successful music generation returns two versions.

Model Selection Guide

Compare the input limits and credit costs below. A newer version is not a guarantee that a particular song will sound better. Listen to your results before choosing what to keep.

The upload limit is a source-audio limit, not a promised output duration. The model selector in each tool shows the options available for that operation.

Core Features

Create music, extend a track, add vocals or an arrangement, make a cover, or separate a song into tracks. Each tool shows its own source requirements and price before submission.

AI Music Generation

Simple starts from an idea. In Custom, use your own lyrics or the lyrics assistant when it is available. Instrumental asks for music without new sung lyrics.

Titles, covers and audio can arrive at different times. A result becomes playable after its audio is ready; you can listen while the other version is still processing. Refreshing does not submit another generation.

1. Choose Simple or Custom, then enter the idea or lyrics.
2. Select a model and review the displayed credits and visibility.
3. Choose Generate once, then follow the saved task on the same page or in Library.
4. Play both versions. Recreate or Use Styles & Lyrics fills a draft; a new generation requires another explicit submission.

Advanced Settings

Available controls include vocal gender, excluded styles, style weight, audio weight and weirdness. The controls shown depend on the tool and model. Start with the displayed defaults and change one setting at a time.

A prompt describes a direction, not a guaranteed performance. Saved lyrics are the submitted text, not a verified transcript or timed lyric display.

Vocal Remover

Choose an eligible generated song from Library. Vocal Remover returns vocal and instrumental tracks; it does not create two alternative songs. Listen to each track before exporting.

Separation can leave traces of other sounds. It does not recover the original recording session or grant rights to third-party recordings.

Stem Splitter

Select an eligible song from Library and review the separation price. The result lists the roles actually returned, such as vocals, drums or guitar. The number and quality of stems can vary.

Use the Stems view in Library to find the outputs. Preview each stem and download the tracks you need for your audio editor.

Add Tracks

Add Vocals starts with an instrumental and lyrics. Add Instrumental starts with a vocal recording and an arrangement description. Choose the matching tab before selecting the source.

Use an owned upload or an eligible Library result. Describe the instruments, mood and vocal direction you want. Compare the two complete returned versions rather than assuming the source was changed exactly as requested.

AI Song Cover

Choose an owned upload or eligible Library track, then describe the new style. Simple keeps the form short; Custom exposes more composition controls.

Listen for changes to melody, phrasing and arrangement. Buying credits does not clear rights in an uploaded recording, and there is no guaranteed artist-voice match.

Extend Song

Upload supported audio or select an eligible Library result. Preview the source and choose Continue At inside the accepted range. Add a style direction and lyrics for the new section, or select Instrumental.

After generation, compare the handoff and ending of both versions. Extend again uses the chosen result as a source; it is another generation with its own displayed credit cost.

WAV Download

Open a song’s More actions or Download Options and choose MP3 or WAV. Both formats use the same download permission. On mobile, open More Actions first, then Download.

WAV export decodes the available audio and saves real PCM samples. Converting an MP3 to WAV does not restore information lost during MP3 compression. Keep the page open while the file is prepared.

A failed export does not delete the saved song. Retry from its result card. Keep a local copy before the displayed file expiry.

Share

Private songs have no public listening link. In Edit Song Detail, choose Public and confirm that the audio, cover, style and lyrics can be viewed by others. Then use Share link.

Set the song back to Private to close its public page and public media access. Your source upload remains private. People who already saved a copy may still have that copy.

Upcoming Features

Personas and playlists are not available. No release date is promised. Use the tools and actions currently shown as available in your workspace.

Best Practices & Workflows

For a new song, describe a clear style and keep the first idea focused. Compare the results before changing one variable. Reusing a draft does not make the next generation free.

For karaoke, generate or select an eligible song, run Vocal Remover, then inspect the instrumental. For a remix, use Stem Splitter and audition the returned tracks before editing them together.

Keep your source unchanged while comparing settings, name results clearly, and save the versions you want before their files expire. One active generation per account prevents conflicting submissions.

Understanding Credits

The selected model determines music-generation credits. Vocal separation and stem splitting have their own fixed prices. Review the quote before submitting; enough credits alone does not prove every paid download or tool permission.

Paid credits do not expire. Public and paid results are retained long term; free private results remain for 90 days and private source uploads for 30 days. Cancelling a subscription takes effect at period end and preserves existing credits.

If credits run out, keep the draft and choose the intended purchase action. A cancelled checkout keeps the draft; a payment return page alone does not unlock a task before payment is confirmed.

Vocal Remover: 60 · Stem Splitter: 315 credits