Describe a song. Get a track. Skrrol's music generator runs ElevenLabs Music and Google Lyria 3 to compose soundtracks, podcast intros, and ad music from a sentence.
Skrrol AI's music generator turns prompts into full-length, original instrumental and vocal tracks. The lineup includes ElevenLabs Music — the same studio behind ElevenLabs voices, now extending into music — and Google Lyria 3, Google's professional-grade music model. Both work the same way from the user's side: write a description of the track you want, pick a duration, hit generate, and the song lands in your library.
The prompt is more than a genre tag. You can describe instrumentation ("warm Rhodes, soft brushes on the drums, upright bass"), tempo ("around 90 BPM"), mood ("hopeful but contemplative"), structure ("intro for 8 bars, then a steady groove"), and reference style ("in the spirit of late-night jazz piano trio recordings"). The more detail, the more reliable the output. For instrumental work, the result is publication-ready; for vocal work, the lyrics and voicing are part of the prompt.
The practical use cases line up with what creators actually need: video soundtracks, podcast intros and bumpers, hold music for ads, scoring for explainer videos, background loops for streams. Because Skrrol runs the music generator next to the video, voice, and image generators, you can produce a complete piece — visuals, voiceover, score — without leaving the studio.
Music generations land directly on a stereo audio track in the editor. From there you can trim, fade, duck under voiceover, EQ, and align beats to cuts. If the track isn't the right length, re-generate with the corrected duration in the prompt — the model can target specific run-times. The Free tier is for testing the prompt language; Standard at €7.99 covers a typical creator's monthly soundtrack needs; Advanced and Advanced Pro are for studios producing per-scene scores at scale.
Run ElevenLabs Music and Google Lyria 3 from one panel. Each has its own strengths — pick per track or A/B both for the same brief.
Describe genre, instruments, tempo, mood, and structure. Re-roll with refined prompts until the brief is hit.
Generate tracks that match a specific run-time — useful for matching a video cut, a podcast intro length, or an ad slot.
Generate purely instrumental beds, or include vocals with prompted lyrics. Skrrol surfaces both modes per model.
Tracks land directly on a stereo audio track. Apply ducking under voice, beat-aligned cuts, fades, and EQ inside the editor.
Outputs are usable on paid Skrrol plans for commercial creator and small-business work, subject to each underlying model's licence terms.
Result: A short, broadcast-ready bumper that sits well under a host's voiceover and ends on a natural beat — drop into your show template.
Result: A swelling instrumental cue that pairs with a montage or a documentary-style intro. Ducks cleanly under narration.
Result: A 30-second bed designed to land beats where ad cuts typically happen — useful for paid social and short-form video.
Result: A seamless background loop suitable for streams, study videos, and YouTube backgrounds where music shouldn't draw attention.
Sign in, click Generate, and pick the Music tab. Choose between ElevenLabs Music and Google Lyria 3.
Cover genre, instrumentation, tempo, mood, structure, and target duration. Be specific — "warm Rhodes piano trio at 90 BPM" beats "jazzy".
Pick a target run-time. The model targets that length so your track fits the video cut or podcast slot you're scoring.
Hit generate. If the first take misses on tempo or feel, refine the prompt and re-roll. Each take is saved as a candidate.
Open the editor, drag the track onto a stereo audio track, and align it with the video. Apply fades and ducking under voice.
Export the full project to MP4, or just the music track as MP3/WAV for separate distribution.
Skrrol AI uses VL credits across all generators — image, video, voice, and music. The same credit pool applies; heavier modalities (video) use more credits per generation than lighter ones (image, voice). Choose a plan and use credits across any generator.
Starter credits to experiment with prompt-driven instrumentals and short loops. Watermarked outputs.
8000 VL credits — soundtracks for weekly videos, short-form background music, podcast bumpers.
17000 VL credits — full-length tracks, multi-section arrangements, and per-scene music variants.
35000 VL credits — composer-level volume for ad campaigns, game scoring, and album-length output.
Outputs are usable for commercial creator and small-business work on paid Skrrol plans, subject to each underlying model's licence. Skrrol surfaces those terms; for high-stakes commercial use, double-check the active model's licence.
Both are professional-grade. ElevenLabs Music tends to feel polished and song-like with strong vocals. Google Lyria 3 leans towards production-ready instrumental work with a wide stylistic range. Skrrol lets you try both per prompt.
Each model has its own duration cap, typically up to a few minutes. For longer pieces, generate sections and stitch them in the editor.
Yes — both models support vocal generation. Specify lyrics and singing style in the prompt. For instrumental-only output, say so explicitly.
Music sits between voice and video in cost. A short bumper is cheap; a multi-minute full song uses more credits. Standard at €7.99 comfortably covers creator-scale music output.
Yes. Trim, fade, EQ, duck under voice, and align beats to video cuts in the editor's audio mixer. If a take is wrong, re-generate and replace it on the timeline.
Every generation opens directly in the Skrrol editor. These features are particularly useful as the next step after a ai music generator — skrrol ai run.
Skrrol AI runs every generator next to a full pro editor. Your work stays on your device. Start free.