Made in @higgsfield_ai PROMPT: Genuine iPhone video, filmed handheld by a laughing friend standing one step above the man, 1 to 2 metres away, vertical. The phone is never still: natural hand shake and micro-jitter in every frame, the friend shifting his weight, bouncing when he laughs, quick reframes to follow the food and the octopus, iPhone autofocus and exposure hunting, slight rolling-shutter wobble, raindrops on the lens. Not a gimbal, not a tripod, not cinematic, no CGI sheen. The video starts exactly on the start image. The camera stays within 2 metres of the man. THE MAN (see references for his face; he wears exactly what the start image shows, plain brown boots): a normal-sized adult, completely natural and alive the whole time, like real candid footage of a friend. Loose shoulders, relaxed posture, constant small natural movements: breathing, blinking, head tilts, hand gestures, shifting on the step, glancing at the friend behind the camera. Real, readable facial expressions that change in real time: bliss, surprise, laughing disbelief. Never frozen, never posing, never stiff. THE OCTOPUS: a real live common octopus (Octopus vulgaris), big market size: wet, slimy, glistening skin, mottled brown-red with small bumps, rows of pale suckers, heavy soft body that sags and flops with gravity, curling arm tips, water streaming off it. Fast, curious, greedy, several arms working at once. CONTINUITY: there is exactly one white bowl in the whole video. Everything that is grabbed stays grabbed; nothing reappears or duplicates. 0-2s: Pure happiness. He takes a long slow sip of wine from the white bowl, eyes closing with pleasure, then lowers it, sighs "mmm" with a huge relaxed grin, wipes his lip with his thumb and laughs at his friend. No octopus yet; only at about 2 s the tip of one wet arm curls up over the edge of the lower step at the bottom of the frame, he does not notice it. 2-6s: The octopus hauls itself up the wet steps from the sea below and three dripping arms raid the board at once: one snatches a paprika-red slice of octopus, one grabs a chunk of bread, one scoops a fistful of green peppers. His grin turns to wide-eyed surprise, he laughs "hey, hey, HEY!", leans in and tries to pull the board back with his free hand. 6-10s: Total mess. The octopus drags the board off the step and flips it: mussels, peppers, bread and red oily octopus slices fly and splatter across the wet granite and onto his jeans and jacket. Another arm flicks the white bowl out of his hand, wine splashing over his chest. He throws both hands up, laughing in disbelief, head shaking, oil and paprika on his fingers. 10-14s: With its loot the octopus slides heavily down the step into the sea with a big wet splash. He leans over the edge on his knees, still laughing, looking down at the ripples, then back at the friend, shoulders shaking. 14-18s: One arm rises from the water and gently sets the empty white bowl back on the step in front of him with a small live green crab inside, legs waving. He stares at the crab, looks straight into the camera with a deadpan face for a beat, then breaks into helpless laughter, wiping paprika off his cheek. The friend's laughter makes the phone bounce. Sound: drizzle, gulls, water lapping on stone, wet slaps of the arms, board clattering on granite, mussel shells rattling, wine splashing, a big splash, the man's "mmm", "hey, hey, HEY!" and laughter, the friend laughing off camera. No music, no subtitles, no text, no logos or brand marks anywhere.

Product: creativemarket.com/Berentine…

Made with: Claude Code Opus 5.5 Extra @claudeai Higgsfield MCP @higgsfield_ai Seedance 2.5, Nano Banana Pro ElevenMusic v2.5 @elevenmusic_io Skills: music-video: my own pipeline for song analysis, lyric timing, cost ledger, and mux verification. seedance-2-5: the house prompt format for every Seedance prompt. built-in-browser: guidance for driving the in-app browser. Local tools: Demucs, faster-whisper, wav2vec2, MediaPipe Face Landmarker, librosa, PyTorch on CUDA and Hugging Face transformers, FFmpeg and ffprobe, yt-dlp, with Node.js, OpenCV, NumPy and Pillow, curl. Python scripts I wrote for this run: prompt builder, assembler, voice map, word-lag sync, per-word residual check, and the B-roll mouth scan.

Product link: creativemarket.com/typemania…