Open editor →

Song in, music video out

Turn a one-line song idea into an original sung track and a music video of a character performing it.

The tin robot dances and mouths along to its own song, with the video driven by the audio · 2026-10-30 · Reported saved run: $0.611. z-ai/glm-5.3-flash · Minimax-Music-2.6 · blackforestlabs/flux-3/klein/text-to-image · music-video-generator

Show audio-driven video. The Avatar node with the Music Video Generator takes a picture and a song and returns a performance timed to the audio. Trim keeps the bill down, since the model bills per second of audio, and Soundtrack puts the full-quality song back.

Saved result

Performer: FLUX.3 Klein 1:1 · reported $0.011 — Song in, music video out
Performer: FLUX.3 Klein 1:1 · reported $0.011 · Download
Song: MiniMax Music 2.6, 33 s returned · reported $0.15 · Download
Music video: Music Video Generator 480p, first 15 s of the song · reported $0.45 · Download

GLM 5.3 Flash writes a two-line verse and a two-line chorus about a red tin-toy robot watering a rooftop garden at sunrise. MiniMax Music 2.6 sings them as bright synth-pop. FLUX.3 Klein paints the robot on its rooftop stage, and the Avatar node sends that picture and the song to the Music Video Generator, which returns a performance with its own scene changes: the robot dances and mouths along while the vocal plays. Trim caps the song at 15 s before the video model, which bills by audio length, and Soundtrack lays the original full-quality song back over the video. Flaws: MiniMax ignored the 15 s duration and returned a 33 s song with an 8 s instrumental intro, so the 15 s cut keeps only the verse and the chorus is not in the video. Move Trim's start to about 7 s to catch the singing. The video was re-encoded for the gallery. Every still and sampled frame was checked and is SFW (a toy robot, no people).

Another run: An owl’s lighthouse lullabySee what the same workflow made with different inputs ↓

Frames, lyrics and audio checked; one real flaw kept · 2026-10-30 · Reported saved run: $0.61104.

Partial. Same graph, a different brief: a slow 72 bpm acoustic folk lullaby sung by a grey owl in a blue scarf on a foggy lighthouse. The lyricist wrote a verse and chorus; the full song (about 52 seconds, although 15 were asked for) is at 72 bpm and all four lyric lines can be heard in it. The owl, its scarf and amber eyes hold across the video, and the beak and wings move. The flaw: the fixed 15-second trim starts at 0 s, so the video carries only the hummed intro and the first two words of the verse. The look also drifts from painterly to more 3D, and a microphone stand appears in the last three seconds. Besides the two text inputs, the song style and video direction prompts were edited; both are listed below.

[Verse] Old grey owl on the foggy shore, Keeps the light forevermore. [Chorus] Sing us home, gentle friend, Guide the ships around the bend.
1. Lyricist: the words · Download
2. MiniMax Music: the whole song · Download
3. FLUX.3 Klein: the performer — An owl’s lighthouse lullaby
3. FLUX.3 Klein: the performer · Download
4. The 15-second music video with the original song · Download
Inputs for this version

Song idea

A slow, gentle acoustic folk lullaby about an old grey owl who keeps a lighthouse on a foggy island and sings the ships safely home at night.

Performer & stage

A wise old grey owl with big amber eyes and fluffy feathers, wearing a small knitted blue scarf, perched on the iron railing at the top of a white and red lighthouse at night, the lighthouse lamp glowing warmly behind it, soft fog over a calm sea, a small sailing boat far away, painterly storybook illustration style, no people, no text.

Song · MiniMax Music 2.6 · prompt

Slow, warm acoustic folk lullaby, 72 bpm, fingerpicked acoustic guitar, soft upright bass, light brushed snare, a gentle breathy lead vocal, cosy foggy night feel.

Music video · Music Video Generator · prompt

Gentle, slow performance music video of the grey owl singing softly on the lighthouse balcony at night, slow drifting camera, sweeping lighthouse beam through the fog, keep the same owl in every shot, no people.

Sampled models: z-ai/glm-5.3-flash, Minimax-Music-2.6, blackforestlabs/flux-3/klein/text-to-image, music-video-generator.

Open this version →Download this versionRun record

Make it yours

  1. Open the graph — save first if the canvas already has your work. Undo brings yours back.
  2. Change Song idea for your own song and Performer & stage for who sings it. Set Trim's start and length to the part of the song you want, since music models often open with an instrumental intro.
  3. Check the editor estimate, then run with your NanoGPT key. GLM lyrics under a cent + FLUX.3 Klein $0.011 + MiniMax Music 2.6 $0.15 + Music Video Generator at 480p, $0.03 per second for 15 s, $0.45. The saved run was $0.611.
  4. Listen to the song first, then check where the singing starts inside the trimmed window. Watch whether the performer moves on the beat and keeps the look of the still.

Download workflow · Original sampled graph · Explore MCP tools · awesome-noodles