AI Music Video Generator

You add a song, choose the part that should play, add photos as references and describe the scene. The song part goes to the AI engine as an audio reference, so the performer sings along and the cuts follow the beat. When the clip is done, your original song replaces the engine's own audio, so the music is exactly your track.

Each generation is up to 30 seconds with Seedance 2.5 or Wan 3.0 Prime and up to 15 seconds with Minimax H3. The song cutter follows the engine you pick: switch engines or video length and the selected part is re-cut from the same starting point.

Seedance 2.5 gives the most cinematic camera work and the most consistent face across cuts. Wan 3.0 Prime also goes to 30 seconds at a much lower credit cost. Minimax H3 is the cheapest and fastest, for 15-second clips such as TikTok and Reels.

Yes. Upload the MP3 or WAV you downloaded, then choose the part you want. You can also make a new song in GoCrazyAI AI Music and pick it from My songs.

The engines time mouth movement and motion to the vocals in the song part you choose. Results are best with a clear lead vocal, a front-facing photo and a prompt that says the person is singing into a microphone. Instrumental parts work well for dance, performance or story clips.

Yes. Add a clear photo of yourself as a reference, or save yourself as a character in Elements so the same face comes back in every clip. Only use photos of people who agreed to it.

It is paid in credits and depends on the engine and length. The exact cost is shown on the Generate button before you start. Minimax H3 is the lowest, then Wan 3.0 Prime, then Seedance 2.5.

Vertical 9:16 for TikTok, Reels and Shorts, horizontal 16:9 for YouTube, plus 1:1 and other ratios depending on the engine. Videos download as MP4 with your song as the soundtrack.