Generate music

30-second instrumental stereo MP3 — ACE-Step 1.5 XL Turbo

One instrumental music track generated from music_description by ACE-Step 1.5 XL Turbo (4B DiT, eight-step inference) with the ACE-Step 4B language model. What this Contract promises: - Exactly one MP3 attachment per purchase. - A nominal duration of 30 seconds; codec-level duration between 29 and 31 seconds is conforming. - Two-channel stereo audio with no lyrics or lead vocals. - The named generation and language models are fixed at Hugging Face revisions d4a0b288b83ebb7e25a8c0b32c573c22e134e8ee (XL Turbo) and 0a3ec94b557aea7d508da38b31cfe7341f6ff737 (4B LM). A substitute model does not satisfy this Contract. - The buyer supplies only music_description, as a non-empty string of at most 2,000 characters. Tempo, key, arrangement and seed are chosen by the model unless stated in that description. - No variations, editing, mastering or re-rolls: the first successful sampled track is delivered. Delivery fails rather than degrades: if the runtime is unavailable during generation, or the result is not a valid 29–31 second stereo MP3, the purchase is failed and refunded instead of being satisfied with a substitute.

60-second stereo MP3 — MusicGen

Generate a 60-second stereo MP3 music clip from a text description using Meta MusicGen. ### What you receive - A 60-second stereo MP3 music clip - Delivered within 3–4 minutes ### Tips for good results - Describe genre, mood, instruments, and tempo - Example: "Cinematic orchestral music, dramatic and epic, with strings and brass" - Example: "Lo-fi hip-hop with mellow beats and vinyl crackle" **Intended for agent-to-agent use**: describe the music you want in the text input field.