Valeria walking through city with GPT Image 2

No es un vlog real. No es una influencer grabándose por la calle. Es GPT Image 2 + Seedance 2.0 en @PolloAIES. Una imagen de referencia, un prompt bien escrito… y Valeria caminando por la ciudad como si fuera contenido real para Reels/TikTok. Prompts completos abajo: https://t.co/8z257ftSyn
- Reply 1 from @ValeriA_Tech
Entré en @itsPolloAI → https://t.co/4JlgkdAIYk usé “Image to Image”, subí una imagen de referencia y le pedí a GPT Image 2 que construyera una hoja de personaje con este prompt: Usa la imagen subida como referencia permanente de identidad para el personaje. La protagonista debe ser exactamente la misma persona de la imagen de referencia. Conserva muy de cerca su identidad: mismo rostro, forma de la cara, proporciones faciales, ojos, cejas, nariz, labios, sonrisa, tono de piel, pelo, textura del pelo y apariencia general reconocible. No inventes una cara nueva. No la embellezcas hasta convertirla en otra persona. No cambies su edad, ni su peinado, ni su identidad. Crea una hoja de referencia visual fotorealista para una creadora de contenido lifestyle que se graba a sí misma por la calle para redes sociales. Esta hoja debe servir como referencia sólida para después generar un vídeo realista estilo vlog. La hoja debe mostrar a la misma protagonista en varios paneles, manteniendo exactamente la misma identidad, el mismo outfit y la misma vibra en todos. Incluye estos paneles: retrato frontal mirando a cámara retrato en ángulo 3/4 perfil lateral sonrisa natural expresión neutra expresión sorprendida mirando hacia arriba mirando hacia un lado sosteniendo la cámara con la mano en ángulo selfie caminando por una calle bonita y cotidiana mientras se graba tipo vlog plano medio enseñando outfit plano espontáneo mirando un escaparate o sosteniendo un café Estilo fotorealista, natural, estética creadora real, vlog, lifestyle, luz natural de día, calle urbana agradable, formato vertical 9:16.

- Reply 2 from @ValeriA_Tech
Después usé Seedance 2.0 utilizando la imagen que me generó GPT image 2 como referencia: Use uploaded image as the permanent character identity, outfit, hairstyle and vibe reference for the entire video. Important: the uploaded image is a reference only. Do NOT show the character sheet, panels, labels, boxes, handwritten notes or collage elements inside the video. The final result must look like real vlog footage, not an animated reference sheet. MAIN CHARACTER: Same exact girl throughout the whole video: Valeria. Preserve her identity very closely in every shot: same face, smile, skin tone, hairstyle, hair texture, facial proportions, eyes, nose, lips and overall recognizable appearance. No face morphing, no identity drift, no hairstyle change, no age change, no outfit change. PERSONALITY: Valeria is natural, spontaneous, cheerful, relaxed and comfortable filming herself. She feels like a real lifestyle creator recording a casual vlog for social media. VIDEO STYLE: Authentic vertical smartphone vlog. Handheld selfie camera, natural walking shake, small wrist movements, imperfect framing, realistic autofocus, slight motion blur, natural daylight, urban street background, casual creator energy. The footage must feel like a real TikTok/Reels clip recorded by Valeria herself while walking through the city. LANGUAGE RULE: All spoken dialogue must be in natural Spanish. No English dialogue. No subtitles. DURATION: Exactly 15 seconds. FORMAT: Vertical 9:16. SCENES: 0:00–0:02 Selfie close-up. Valeria is walking down a nice city street holding the camera with her own hand. Slight handheld shake from her steps. She looks into the lens, smiles naturally and says: “Hoy he salido a grabar un poco por la ciudad…” 0:02–0:04 She turns the camera slightly to the side to show the street, a café terrace or a pretty shop window. The movement must feel handheld, with a small imperfect tilt, like a real vlog. 0:04–0:06 The camera comes back to her face. She laughs softly, keeps walking and says: “Me encanta cuando la luz queda así.” 0:06–0:08 She briefly adjusts her hair with her free hand while still walking. Natural blinking, soft smile, slight wind movement in the hair. 0:08–0:10 The camera tilts down a little to show part of her outfit and her steps walking on the sidewalk, then tilts back up toward her face. 0:10–0:12 She looks toward a shop window or something off-camera, reacts with a small surprised smile and turns the camera slightly to show it for a moment. 0:12–0:15 The camera returns to her face. She smiles, walks closer to the camera and says naturally: “Vale, luego os enseño lo que he grabado.” End like a real social media clip.


Upload your character and run this pipeline.
The prompts above are pre-filled in the workbench. Swap in your image and tweak the topic — first generation is free.
Try this technique →