ShotTide中文Sound for your story
Start with a scene, a lyric, or your own track. Build a soundtrack around it.
Loading saved audio…
Generate a track to listen, compare and download. Each take keeps its own settings.
An AI music generator turns a description into a musical performance. Use instrumental mode for narration beds, supply your own lyrics for a song, or let the model develop lyrics from an idea. Reference mode rearranges an authorized song with singing while retaining its melody; instrumental references are unsupported.
Describe the mood, instruments and energy of the scene. For a voiceover bed, request restrained drums and a simple arrangement. Instrumental mode explicitly asks for no vocals; always listen to the result before mixing it under speech.
Organize your own words with [Verse], [Chorus] and [Bridge] tags. Describe tempo, key and vocal direction in the style field. If you have an idea rather than lyrics, choose the automatic songwriting mode.
Upload an authorized 20-120 second song excerpt with singing, or select and preview a vocal excerpt from a longer track here. Instrumental references are unsupported. Choose a new style and optionally replace the lyrics. This uses Music Cover to retain the melody while changing the arrangement. It does not clone the source singer.
Pick instrumental, your lyrics, automatic songwriting or reference arrangement. The input fields follow the mode.
Use genre, mood, instruments, BPM and key. Edit a preset to suit your scene rather than submitting an empty brief.
Review the price, generate, then listen to the whole track. Download MP3 or continue into Beat Sync for a video edit.
Yes. Instrumental mode uses MiniMax Music 2.6 with its instrumental setting enabled. Describe the sound in the style field; lyrics are not sent in this mode.
Yes. The lyrics mode accepts up to 3,500 characters. Use section tags and line breaks to guide the arrangement. English and Mandarin are useful starting points; review pronunciation in every generated result.
Original songs and instrumental tracks cost 3 credits per generation. A reference arrangement costs 5 credits. Downloading an existing result does not create another generation charge.
A length written in a prompt is guidance, not an exact duration guarantee. The model generates a complete track. Listen to the result and select a section for your video edit.
Yes, in reference arrangement mode. Upload a 20-120 second song excerpt that contains singing and that you own or have permission to adapt. Purely instrumental references are unsupported. For longer sources up to 10 minutes and 30 MB, select, preview and apply an excerpt in your browser before generating. The model preserves melody while changing style; it does not provide clearance for someone else’s recording or composition.
No blanket guarantee is made. You remain responsible for rights to lyrics, reference recordings and other inputs. Check the relevant service terms and your distribution requirements before commercial release.