Introduction
A That Seoul Energy GRWM workflow is an AI pipeline that produces a vertical get ready with me video in the clean minimal K aesthetic, complete with soft K pop instrumental and an energetic female voiceover. This guide shows the seven node Vertex build, the prompts you need, and a hero image prompt you can run right now.
What Is the That Seoul Energy Aesthetic
The That Seoul Energy aesthetic is clean minimal and bright. Think pale oak floors, white walls, glass skin makeup, oversized pastel knitwear, dried florals in clear glass, a single matcha or iced americano prop. The color palette skews cool cream pink and neutral. Reference points are Dazed Korea, Vogue Korea editorial, and contemporary Seoul street style. The mood is calm confident aspirational.
What the Workflow Produces
A 12 to 18 second vertical 9:16 video with clean editorial framing, a subtle camera movement, a bright Korean or English voiceover, and a soft K pop instrumental bed. The output works for TikTok, Reels, and YouTube Shorts without additional post.
Hero Image Prompt You Can Run Today
Run this prompt directly in Flux Kontext, Midjourney v7, or Nano Banana to generate the hero image that accompanies this blog post. The prompt is tuned for a 9:16 vertical composition and can be adapted to 1:1 for feed posts.
A young Korean woman in her early twenties, porcelain glass skin with a dewy finish, straight medium length dark brown hair with soft face framing layers, straight natural brows, glossy peach lips, a single small pearl earring. She wears an oversized cream cable knit cardigan over a white camisole. She stands in a bright minimal Seoul studio apartment with white walls, a pale oak floor, a single dried eucalyptus branch in a clear glass vase, a small ceramic mug of iced matcha on a round marble side table. Soft diffused window light from camera left fills the scene evenly. Shot on 50mm lens, clean digital capture, cool neutral color grade with gentle pink undertones, shallow depth of field at f 2.2, minimalist K editorial mood similar to Dazed Korea photography. Vertical 9:16 composition.
Step by Step Vertex Build
Seven nodes: interface form, CHAT visual prompt writer, Flux Kontext keyframe, Veo animation, TEXT_TO_SPEECH with Korean or English voice, music generator with K pop instrumental prompt, and a final VIDEO_AUDIO_MERGE.
Node Build Table
Build these nodes in order on the Vertex canvas. Each node below lists its execution mode and key fields with wiring details.
Step 1: node_interface_form
Execution Mode: INTERFACE_FORM
Fields and Wiring: node_type INTERFACE. Outputs: character_detail (TEXT), setting_detail (TEXT), voiceover_script (TEXTAREA), voice_language (SELECT: korean english), aspect_ratio (ASPECT_RATIO default NINE_SIXTEEN)
Step 2: node_chat_image_prompt
Execution Mode: CHAT
Fields and Wiring: System writes Flux image prompt from character_detail and setting_detail plus anchor style language. Output: generated_text
Step 3: node_flux_seoul_keyframe
Execution Mode: CREATE
Fields and Wiring: Inputs: prompt from chat node, aspect_ratio from interface. Output: generated_asset IMAGE
Step 4: node_veo_drift
Execution Mode: VIDEO
Fields and Wiring: Inputs: reference_image_1 from flux, prompt gentle static frame with slight camera rise over 8 seconds, aspect_ratio NINE_SIXTEEN. Output: generated_asset VIDEO
Step 5: node_tts_k_voice
Execution Mode: TEXT_TO_SPEECH
Fields and Wiring: Inputs: text_input from interface voiceover_script. voice_id switches based on voice_language selection. Output: generated_audio SPEECH
Step 6: node_music_kpop
Execution Mode: MUSIC_GENERATOR
Fields and Wiring: Prompt: soft K pop instrumental dreamy synth pads light percussion 95 BPM feminine bright Seoul morning 15 seconds. Output: generated_audio AUDIO
Step 7: node_final_merge
Execution Mode: VIDEO_AUDIO_MERGE
Fields and Wiring: Inputs: video_reference from veo, audio_reference from a prior compose step combining TTS and music. Output: merged_video
Flux Keyframe Prompt
Anchor prompt for node_flux_seoul_keyframe: young Korean woman early twenties, porcelain glass skin dewy finish, straight medium dark brown hair face framing layers, glossy peach lip small pearl earring, oversized cream cable knit cardigan over white camisole, bright minimal Seoul studio apartment, pale oak floor white walls, dried eucalyptus in clear glass vase, ceramic mug of iced matcha on marble side table, soft diffused window light from left, clean digital capture cool neutral grade with pink undertones, 50mm lens f 2.2, Dazed Korea editorial mood, vertical 9:16.
Veo Motion Prompt
Motion prompt for node_veo_drift: subject holds relaxed neutral pose, gentle camera rise from waist level to eye level over eight seconds, soft window light remains consistent, no cuts, slight sway of cardigan fabric, single sustained take, clean digital motion without grain.
FAQ
Does this workflow support Korean voiceover?
Yes. Set the voice_language field to korean on the interface form and select a Korean voice ID in the TEXT_TO_SPEECH node. MiniMax is recommended for Korean pronunciation fidelity if ElevenLabs accuracy is insufficient.
Can I add on screen text captions?
Yes. Add a final node in Vertex that handles caption overlay or export the merged video and add captions in CapCut or Premiere. Native caption generation is on the Vertex roadmap.
How do I keep the same character across multiple videos?
Save the first successful Flux keyframe as an anchor image. In subsequent workflows, use EDIT mode with the anchor as source_image rather than CREATE mode. This preserves facial identity across a series.
