Introdução
📌 Ferramentas de Inteligência Artificial para os Alunos
⭐ FERRAMENTAS OBRIGATÓRIAS
ChatGPT
🧠 O que é: Assistente de IA para criação de textos, ideias, roteiros, prompts, estudos, planejamento e suporte geral.
💡 Ponto forte: Extremamente versátil. Serve para praticamente tudo: aprender, criar, organizar e estruturar qualquer tipo de conteúdo.
Manus AI
🔗 https://manus.im/invitation/NZENMNIBCWWQQV
🧠 O que é: Um agente de IA autônomo capaz de executar tarefas completas, tomar decisões e automatizar processos, não apenas responder perguntas.
💡 Ponto forte: Atua como um “funcionário digital”, ideal para automações, pesquisas profundas e execução de tarefas mais complexas e em grandes quantidades.
Nordy AI
🧠 O que é: Plataforma que permite usar ComfyUI direto no navegador, sem instalar nada e sem precisar de computador potente ou placa de vídeo.
💡 Ponto forte: Acesso fácil a workflows avançados de geração de imagens com IA, rodando 100% na nuvem e sem filtros.
CivitAI
🧠 O que é: uma plataforma online e comunidade colaborativa focada em compartilhar, descobrir e baixar modelos de inteligência artificial, principalmente voltados para geração de imagens com IA usando tecnologias como Stable Diffusion, LoRAs, embeddings e checkpoints.
💡 Ponto forte: reúne milhares de modelos criados por usuários, com feedback, avaliações e exemplos visuais — o que facilita encontrar e usar modelos específicos para estilos artísticos ou projetos criativos de IA.
🔹 FERRAMENTAS OPCIONAIS (Complementares)
Google Gemini
🔗 https://gemini.google.com/
🧠 O que é: IA do Google para criação de textos, ideias, imagens e auxílio em tarefas do dia a dia.
💡 Ponto forte: Integração com Google Docs, Gmail e Drive.
Google Flow (Labs)
🔗 https://labs.google/fx/pt/tools/flow/
🧠 O que é: Ferramenta experimental do Google para criação de projetos criativos com IA, combinando lógica, texto e visual.
💡 Ponto forte: Ideal para testar ideias e protótipos sem precisar programar.
Perplexity AI
🧠 O que é: IA focada em pesquisa e respostas com fontes verificáveis.
💡 Ponto forte: Excelente para estudos, pesquisas e validação de informações.
NotebookLM (Google)
🧠 O que é: IA que analisa PDFs, textos e documentos enviados pelo usuário e responde com base nesse material.
💡 Ponto forte: Ótima para estudar conteúdos longos e transformar documentos em resumos e insights.
Microsoft Copilot
🔗 https://copilot.microsoft.com/
🧠 O que é: Assistente de IA da Microsoft para produtividade, escrita e apoio em tarefas profissionais.
💡 Ponto forte: Integração com Word, Excel, PowerPoint e Windows.
Bônus
Bônus 01: Movimentos de Câmera
Nesta aula, você vai aprender a controlar movimentos reais de câmera com IA, como órbita, subida de câmera e zoom extremo, mantendo consistência total da influencer e do cenário.
Os prompts foram criados para evitar efeitos falsos (slide ou zoom artificial) e simular comportamento cinematográfico de câmera real.
Use os campos abertos para adaptar a fala e o estilo ao seu projeto.
1 – Órbita parcial (arco / 360° controlado),
POSITIVE PROMPT (ENGLISH)
Create a hyper-realistic video using a true 3D camera system.
The subject is a single digital influencer with absolute identity lock: same face geometry, facial proportions, body shape, skin tone, hair, makeup, clothing, and visual identity across all frames. No morphing, no drift, no regeneration. The environment must be fully three-dimensional, physically coherent, and spatially stable. Background elements must exist at different depths (foreground, midground, background) to enable visible parallax during camera movement. Lighting must remain fixed and physically consistent.
CAMERA MOVEMENT — CRITICAL
The camera must execute a real spatial orbit movement, not a 2D pan, not an image slide.
Simulate a physical camera rotating around the subject on a circular path, completing exactly 90 degrees (¼ of a full 360° orbit).
Technical constraints:
The camera rotates around the vertical Y-axis
The camera maintains constant distance from the model
The model stays fixed in world space
The background shifts relative to the subject with clear parallax
Perspective must visibly change during motion
The movement must resemble a cinematic arc shot / orbital tracking shot, as if filmed on a dolly or robotic arm.
❌ The frame must NOT translate sideways
❌ The image must NOT slide
❌ The subject must NOT move with the frame
✅ Only the camera rotates in 3D space
The orbit starts front-facing and ends at a three-quarter profile angle, rotating smoothly to the left, then stopping
The influencer speaks naturally during the orbit, with:
Perfect lip synchronization
Realistic micro-expressions
Stable eye tracking toward the moving camera
Dialogue spoken by the influencer:
[INSERT THE LINE HERE]Visual style: ultra-photorealistic, cinematic realism, natural depth of field, physically accurate perspective, sharp focus, realistic skin texture.
NEGATIVE PROMPT (ENGLISH — STRICT)
2D motion, lateral sliding, image translation, flat pan, fake camera movement, no parallax, background moving with subject, subject locked to frame, slideshow effect, frame shifting, zoom simulation, camera teleporting, cuts, transitions, multiple cameras, identity drift, face reshaping, lighting flicker, background replacement, CGI look, render look, cartoon, anime, illustration, plastic skin, blur, noise, artifacts, robotic animation.
⚠️ IMPORTANTE (honesto e direto)
Se mesmo com esse prompt ainda virar slide, isso significa que:
O modelo de vídeo não suporta câmera orbital real
Ele só simula movimento 2D
Nesse caso, a única saída real é:
Usar Veo 3 (camera motion enabled)
Ou Runway Gen-3 com “camera orbit / arc”
Ou gerar frames separados + interpolação
2 – Zoom progressivo até o olho da influencer
POSITIVE PROMPT (ENGLISH — EXTREME EYE ZOOM)
Create a hyper-realistic video with a single digital influencer and absolute identity lock. The same face geometry, eye shape, iris pattern, skin texture, body proportions, hair, makeup, and visual identity must remain perfectly consistent throughout all frames, with no morphing, no regeneration, and no identity drift. The environment must remain fully stable and unchanged, with consistent lighting, color temperature, and spatial coherence. No background swaps, no lighting flicker, no object movement. Lighting must be cinematic and soft, with realistic reflections and catchlights in the eyes.
CAMERA MOVEMENT — CRITICAL (EYE ZOOM)
The camera must perform a slow, continuous, cinematic zoom-in toward the influencer’s eye, simulating a real optical camera zoom, not a digital crop.
Technical constraints:
The zoom starts from a medium or close-up framing
The camera gradually moves closer, ending in an extreme close-up of a single eye
Eye details must be sharp and realistic (iris texture, reflections, moisture)
Perspective must remain natural during the zoom
No sudden acceleration or deceleration
The motion must feel physically controlled, like a professional cinema lens
❌ This must NOT be a cut
❌ This must NOT be a jump zoom
❌ This must NOT be a digital crop or scale
❌ This must NOT be a slideshow transition
The influencer speaks naturally during the zoom, with:
Precise lip sync before the mouth exits frame
Subtle facial micro-expressions
Natural eye movement without distortion
Stable gaze aligned with the camera
Dialogue spoken by the influencer:
[INSERT THE LINE HERE]
Visual style: ultra-photorealistic, cinematic realism, macro-level eye detail, natural depth of field, shallow focus near the eye, realistic skin pores, professional lighting.
NEGATIVE PROMPT (ENGLISH — STRICT)
Jump cuts, instant zoom, digital crop, image scaling, 2D zoom effect, slideshow motion, fake camera zoom, frame popping, distortion of the eye, warped iris, duplicated pupils, asymmetric eyes, flickering highlights, CGI look, render look, cartoon style, anime style, illustration, plastic skin, blur, noise, artifacts, robotic motion, exaggerated expressions, lighting changes, background replacement, identity drift, face reshaping.
DICA IMPORTANTE (realista)
Nem toda IA consegue zoom óptico real até macro do olho. As que costumam funcionar melhor são modelos que entendem lens behavior / macro close-up.
3 – Movimento vertical (tilt up)
POSITIVE PROMPT (ENGLISH — DETAILED CAMERA CONTROL)
Create a hyper-realistic video using a true physical camera movement, not a digital zoom or 2D slide. The subject is a single digital influencer with absolute identity lock: identical face structure, body proportions, skin tone, hair, makeup, clothing, and overall appearance across all frames, with no morphing, no drift, and no regeneration. The environment must be fully three-dimensional and spatially coherent, with clear foreground, midground, and background layers to ensure visible parallax during camera movement. The scene must remain completely stable: no background changes, no object movement, no lighting variation. Lighting should be natural and cinematic, physically accurate, with stable shadows and realistic depth.
CAMERA MOVEMENT — CRITICAL (VERTICAL)
The camera must perform a real vertical camera movement (tilt up), simulating a physical camera mounted on a rig:
The camera starts at a lower framing position
The camera moves upward in 3D space, tilting up smoothly
The subject remains centered in frame
The camera maintains constant distance from the subject
Perspective must change naturally during the movement
Foreground and background elements must shift relative to the subject, showing real parallax
❌ This must NOT be a digital zoom
❌ This must NOT be a sliding image
❌ The subject must NOT move with the frame
✅ Only the camera physically moves upward
The movement should be smooth, continuous, cinematic, and realistic, as if operated by a professional cinematographer using a dolly or crane. The camera motion starts slowly, rises steadily, and stops naturally at the final framing. The influencer speaks naturally during the movement, with:
Accurate lip synchronization
Subtle, realistic facial expressions
Stable eye tracking toward the camera
Dialogue spoken by the influencer:
[INSERT THE LINE HERE]Visual style: ultra-photorealistic, cinematic realism, physically accurate perspective, natural depth of field, sharp focus, realistic skin texture, professional lighting.
NEGATIVE PROMPT (ENGLISH — STRICT)
Digital zoom, 2D slide, flat vertical motion, fake camera movement, no parallax, subject moving with the frame, background locked to subject, slideshow effect, frame translation, pan simulation, jump cuts, transitions, camera teleporting, multiple angles, identity drift, face reshaping, body distortion, lighting flicker, background replacement, CGI look, render look, cartoon style, anime style, illustration, plastic skin, blur, noise, artifacts, robotic motion, exaggerated expressions, clothing changes, hair changes.
DICA FINAL (importante)
Se mesmo com esse prompt ainda virar “zoom”, então a IA não suporta tilt vertical físico real — apenas simula. Nesse caso, o melhor resultado vem usando crane / dolly up em ferramentas que aceitam instruções de câmera física.
Bônus 02: Zoom Inteligente
Guia Técnico: Criação de Efeito de Zoom Inteligente com Inteligência Artificial
(1) Fundamentos e Ferramentas
Categoria | Ferramentas Recomendadas | Funcionalidade Principal |
Geração de Imagem | Midjourney, Stable Diffusion, Adobe Firefly | Criação da sequência de imagens com outpainting. |
Geração de Vídeo | Luma Dream Machine, Kling AI, Runway Gen-2 | Automatização da transição entre frames. |
Edição e Scripts | Adobe Premiere, After Effects, Python (OpenCV) | Alinhamento manual ou programático dos frames. |
(2) Procedimento para Imagens e Outpainting
Etapas do Fluxo de Trabalho:
(3) Produção de Vídeo e Animação
Abordagem Automatizada
Abordagem Manual e Programática
(4) Engenharia de Prompts para Zoom Infinito
Prompt Positivo (Positive Prompt)
Prompt Negativo (Negative Prompt)
(5) Considerações Técnicas e Melhores Práticas
Bônus 03: Efeitos Cinematográficos
Guia Técnico: Domínio de Efeitos Cinematográficos com Inteligência Artificial
(1) O Triângulo da Cinematografia Digital
Iluminação (Lighting)
Composição e Enquadramento (Composition)
(2) Dinâmica de Movimento em Vídeos de IA
Movimento de Câmera | Comando de Prompt (Inglês) | Objetivo Narrativo |
Dolly In (Push) | CAMERA: SLOW DOLLY IN | Aumentar a tensão ou focar na reação emocional do personagem. |
Dolly Zoom | CAMERA: DOLLY ZOOM | Criar uma sensação de desorientação ou revelação chocante (Efeito Vertigo). |
Orbit 360 | CAMERA: 360 ORBIT | Estabelecer a grandiosidade de um cenário ou a importância de um objeto. |
Pedestal Up/Down | CAMERA: PEDESTAL UP | Revelar gradualmente elementos da cena de baixo para cima. |
(3) Engenharia de Prompts Cinematográficos Abaixo, apresentam-se modelos de prompts otimizados para obter resultados de alta fidelidade visual. Prompt Positivo (Positive Prompt) Prompt: “Cinematic wide shot of [SUBJECT], shot on anamorphic lens, 35mm film grain, Rembrandt lighting, deep depth of field, teal and orange color grading, highly detailed textures, volumetric fog, masterpiece, 8k, directed by [STYLE OF DIRECTOR], sharp focus, cinematic atmosphere.” Prompt Negativo (Negative Prompt) Negative Prompt: “cartoon, 3d render, anime, low resolution, blurry, overexposed, flat lighting, watermark, text, signature, distorted faces, bad anatomy, plastic skin, oversaturated colors, low quality, flickering, morphing artifacts.” (4) Dicas de Finalização e Estilo Color Grading: Utilize termos como “Teal and Orange” para um visual de blockbuster moderno ou “Monochromatic Noir” para um estilo clássico. Textura de Filme: Adicionar “Kodak Portra 400” ou “Fujifilm Velvia” ajuda a IA a simular a granulação e a resposta de cor de filmes analógicos reais . Profundidade de Campo: O comando “Shallow depth of field” é crucial para isolar o sujeito e criar um visual profissional.
Bônus 04: Transições que Prendem a Atenção
Guia Técnico: Transições de Alto Impacto com Inteligência Artificial
(1) Categorias de Transições que Capturam a Atenção
A. Match Cut (Corte de Correspondência)
B. Morphing (Metamorfose)
C. Seamless Transitions (Transições Invisíveis)
(2) Fluxo de Trabalho Otimizado com Inteligência Artificial
(3) Prompts Especializados para Transições Fluidas
Prompt para Morphing (Definição de Início e Fim)
Start Frame Prompt: “A detailed close-up of a burning candle, flickering flame, cinematic lighting, shallow depth of field, warm tones.” End Frame Prompt: “A majestic lighthouse at night, beam of light rotating, stormy sea background, dramatic lighting, cool tones, long exposure.” Transition Instruction: “Seamless morphing transition from the candle flame into the lighthouse beam, maintaining central light position, smooth transformation, dreamlike quality.”
Prompt para Zoom Through (Transição de Mergulho)
(4) Dicas Essenciais para Retenção de Audiência
Bônus 05: Texturas de Pele Realistas
Guia Técnico: Consistência de Textura de Pele Realista em Múltiplas Gerações com IA
(1) Fundamentos da Consistência: Referências e Parâmetros
A. Character Reference (–cref) no Midjourney
B. Uso de Seeds (Sementes)
(2) Padronização de Prompts: O “DNA” da Pele
O Bloco de Consistência (Exemplo):
(3) Controle de Iluminação e Ambiente para Consistência
Elemento | Recomendação para Consistência | Justificativa |
Lente | Mantenha a mesma distância focal (ex: 85mm, 50mm). | Lentes diferentes alteram a perspectiva e a compressão, afetando a percepção da textura da pele. |
Iluminação | Defina um tipo de luz base (ex: “soft box lighting”, “natural daylight”). | A qualidade e direção da luz afetam diretamente a visibilidade de poros, brilhos especulares e o efeito de subsurface scattering. |
Estilo de Renderização | Utilize –style raw (no Midjourney) ou termos como “unfiltered photo”. | Isso instrui a IA a evitar a aplicação de filtros de embelezamento ou estilização automática que podem suavizar ou alterar a textura da pele. |
Ajuste Fino com Variações: Se pequenas inconsistências surgirem, utilize as ferramentas de variação da plataforma de IA para refinar a imagem, mantendo a seed e o –cref.
(5) Exemplo de Prompt Consistente
Bônus 06: Coleção de Prompts Prontos
PROMPT 1 – Criação da influencer
POSITIVE PROMPT (CONTINUOUS TEXT)
Create an extremely realistic photographic image of an adult sensual woman, approximately 25 years old, designed for a male audience and for Instagram Reels, using a strong UGC (User Generated Content) aesthetic, giving the impression of a real frame captured from a spontaneous video. The model must be blonde, very attractive in a natural and realistic way, with authentic human features and no exaggerated beauty standards. Sensuality should be subtle, elegant, and natural, expressed through posture, gaze, and presence, never through nudity or explicit content. The skin must show true human texture, with visible pores, micro-imperfections, slight natural shine, and subtle tone variations. The face should display a confident, engaging expression, with a natural gaze and a soft smile or casual look, never forced. Blonde hair must look realistic, with visible individual strands, natural volume, and slight movement. Makeup should be light and realistic, enhancing natural facial features without artificial effects. Clothing must be modern, feminine, and discreetly sensual, such as a fitted dress, elegant top, or stylish casual outfit, with realistic fabric and fit. Body proportions must be natural, with a relaxed yet confident posture. Lighting should be soft and natural, like window light or ambient daylight, with realistic shadows and no studio look. The environment should be simple and real, such as a bedroom, living room, neutral wall, or balcony, reinforcing the UGC feel and avoiding any campaign or editorial appearance. The framing must be vertical 9:16, with realistic depth of field, sharp focus, DSLR-quality photography, natural slightly warm colors, and an overall human, authentic, attractive, and trustworthy look, like a perfect still frame from a real Reel.
NEGATIVE PROMPT (CONTINUOUS TEXT)
Completely avoid nudity, explicit content, as well as vulgar or unnatural poses. Do not generate doll-like, CGI, illustrated, anime, cartoon, or 3D appearances. Avoid plastic, overly smooth skin or heavy beauty filters. Do not create overly symmetrical faces, unrealistic beauty, or artificial perfection. Avoid harsh studio lighting, cinematic lighting effects, or unrealistic shadows. Do not use heavy makeup, blur, skin smoothing, or artificial enhancement filters. Avoid transparent, fetishized, or unrealistic clothing. Do not generate distorted body proportions, unnatural poses, or incorrect anatomy. Avoid deformed hands, extra fingers, anatomical errors, or rendering glitches. Do not use artificial or obviously AI-generated backgrounds. Avoid low resolution, lack of sharpness, or a corporate stock photo look.
PROMPT 2 – Close Up de uma imagem
POSITIVE PROMPT (CONTINUOUS TEXT)
Use the uploaded reference image as an absolute and unchangeable base. Generate 10 ultra-realistic close-up images, where each image must contain ONLY ONE unique close-up, filling 100% of the frame, with no splits, collages, grids, multiple panels, or side-by-side comparisons. All images must feature the exact same model from the original image, preserving identical facial identity, apparent age, skin tone and texture, makeup, hair, clothing, and environment, with no aesthetic changes. All images must be created in the same location as the reference photo, with identical and consistent lighting, faithfully matching the original light direction, intensity, shadows, and colors. Each generation must present one exclusive close-up, varying head angle, face tilt, and eye direction, creating 10 distinct and natural poses, like sequential frames from the same real video. The close-ups must include, in separate and unique images: mouth, one eye, the other eye, both eyes together, nose, ear, and additional extremely close facial views, always one main focal area per image, totaling 10 distinct images, all wearing the same outfit and in the same environment. Skin must display true human texture, with visible pores, micro-imperfections, and natural shine. Eyes must show realistic depth, natural reflections, and precise focus. Lips must retain natural texture with no artificial smoothing. Visible hair must show individual strands naturally. Framing must be extremely close yet optically realistic, with no lens distortion, professional DSLR or mirrorless quality, realistic depth of field, and extreme sharpness. Premium UGC style, fully human, spontaneous, and absolutely faithful to the reference image.
NEGATIVE PROMPT (CONTINUOUS TEXT)
Do not generate split images, collages, grids, multi-frame layouts, or side-by-side comparisons. Do not alter facial identity, face shape, age, makeup, hair, clothing, environment, or lighting. Do not repeat identical close-ups without real angle variation. Avoid any CGI, illustration, anime, cartoon, or 3D appearance. Do not apply beauty filters, blur, skin smoothing, or texture reduction. Avoid lens distortion, unrealistic framing, or incorrect proportions. Do not generate artificial eyes, exaggerated shine, or plastic skin. Avoid color shifts, skin tone changes, or visual inconsistencies between images. Do not create anatomical errors, extra parts, or deformations. Avoid low resolution, lack of sharpness, excessive noise, or stock photo appearance.
PROMPT 3 – Modelo da imagem em lugares e roupas diferentes
POSITIVE PROMPT (CONTINUOUS TEXT)
Create 10 ultra-realistic photographic images, each image being one single shot (no splits or collages), featuring the same adult female model, approximately 25 years old, blonde, naturally beautiful, with a subtle, elegant sensuality, designed for a male audience and optimized for Instagram Reels (9:16) with a strong UGC (User Generated Content) aesthetic. The model must look fully human and realistic, with authentic facial identity preserved across all images, while changing locations and outfits in every image. Sensuality should be conveyed through posture, gaze, and presence—never explicit, never nude. Each image must take place in a different real-world location, such as: a softly lit bedroom, a modern living room, a balcony at golden hour, an urban café, a minimalist hallway, a sunlit window corner, a rooftop at sunset, a neutral textured wall, a hotel room, and an outdoor city street. Every image must feature a different outfit, all modern, feminine, and discreetly sensual, such as fitted dresses, elegant tops, casual chic outfits, stylish sets, or minimal fashion looks, with realistic fabric behavior and natural fit. The model should display natural micro-poses and variations in body angle, head tilt, shoulder position, and eye direction, making each image feel like a different frame from real UGC videos, never repetitive. Skin must show true human texture, visible pores, micro-imperfections, subtle shine, and realistic color variation. Hair must show individual strands, natural volume, and believable movement. Makeup should be light and realistic, enhancing natural features without artificial effects. Lighting must be natural and environment-appropriate (window light, ambient daylight, soft indoor lighting), with realistic shadows and no studio appearance. Use professional photographic realism with DSLR or mirrorless quality, realistic depth of field, sharp focus, and natural, slightly warm tones. Each image must feel authentic, spontaneous, attractive, and trustworthy—like real content captured casually, not a staged campaign.
NEGATIVE PROMPT (CONTINUOUS TEXT)
Avoid nudity, explicit sexual content, or exaggerated sexualization. Do not generate split images, collages, grids, or multiple frames in a single image. Avoid CGI, illustration, anime, cartoon, or 3D styles. Do not create plastic, overly smooth skin or apply beauty filters, blur, or skin smoothing. Avoid unrealistic beauty standards, overly symmetrical faces, or artificial perfection. Do not use harsh studio lighting, cinematic lighting, or unrealistic shadows. Avoid vulgar poses, forced expressions, or unnatural body positions. Do not repeat outfits or locations. Avoid distorted proportions, lens distortion, anatomical errors, deformed hands, or extra fingers. Do not change the model’s identity between images. Avoid artificial or obviously AI-generated backgrounds. Do not produce low resolution, lack of sharpness, heavy noise, or corporate stock photo appearance.
PROMPT 4 – Imagem artística de sombras
ENGLISH — HYPER-DETAILED PROMPT (CONTINUOUS TEXT) Create a hyper-realistic photographic image capturing a single static outdoor location where the passage of time is visually represented through shadows. The environment must remain exactly the same, but multiple shadows appear on the ground and walls, each one corresponding to a different moment of the day (early morning, midday, late afternoon, sunset). The shadows must be physically accurate, following correct sun angles, direction, length, softness, and intensity for each time period. Lighting behavior must obey real-world physics, with realistic ambient light, bounce light, and natural color temperature shifts throughout the day. Textures of surfaces (concrete, stone, wood, pavement, walls) must show fine detail, wear, imperfections, and realistic material response to light. The scene must feel completely natural and believable, as if the camera captured a rare visual phenomenon in real life. Use professional photographic realism, correct perspective, natural depth of field, sharp focus, and true-to-life color reproduction. No surreal elements, no fantasy, no illustration — the image must look like a real photograph capturing an extraordinary but physically possible moment.
PROMPT 5 – Texto narrativo curto extremamente imersivo
PROMPT 6 – Copiar a estrutura visual de uma imagem
Using the reference image only as a structural and stylistic guide, create a completely new image that preserves the same pose, body positioning, camera angle, framing, and overall visual composition, while fully changing everything else. The person must be entirely different from the reference: different face, body, age, ethnicity, hair, and identity. The environment must be completely new, set in a different location with distinct surroundings, textures, and atmosphere. The clothing must be entirely different, with a new style, colors, fabrics, and fit.
Maintain the same visual style and mood of the original image (such as lighting softness, contrast level, depth of field, and aesthetic tone), but ensure there is no visual similarity that suggests copying. The image must feel inspired by the structure, not replicated. Lighting should adapt naturally to the new environment while respecting the original style. The composition must mirror the original pose and framing precisely, but the subject, setting, and details must be fully original. The final image should be highly realistic, with natural human proportions, realistic textures, and believable lighting. Avoid any direct resemblance to the original model, outfit, or background. The result must look like a new creative work, sharing only the underlying pose and stylistic structure, never the content itself.
PROMPT 7 – Criar o roteiro completo de um vídeo de historinha infantil
Bônus 07: Influencers de IA Prontas
Para ter acesso a mais de 100 influencers já prontas, acesse o link do Drive abaixo:
https://drive.google.com/drive/folders/1F6GiQ4UVFYD6J3SQWSaTU0gH609BauHx?usp=drive_link
Módulo 01 - Utilização das IAs
Aula Nº 1 - Geração de Imagens
Aula Nº 1 - Atualização (05_02_2026)
Aula Nº 2 - Geração de Videos
Aula Nº 3 - Prompts
Texto base para geração de prompt:
Você é um especialista em criação de prompts para inteligência artificial. Crie um prompt completo, claro e estratégico para o seguinte objetivo: [DESCREVA AQUI O QUE DEVE SER CRIADO]. Esse conteúdo deve ser feito para [DESCREVA O PÚBLICO-ALVO] e será utilizado em [INFORME ONDE SERÁ USADO – ex: Instagram, Reels, Stories, anúncio, vídeo, aula, imagem, etc]. O estilo e o tom da linguagem devem ser [DEFINA O TOM – ex: persuasivo, simples, profissional, motivador, UGC, educativo, etc]. O formato do resultado deve ser [DEFINA O FORMATO – curto ou longo, em tópicos ou parágrafos, com ou sem emojis, com ou sem CTA]. Considere também as seguintes regras e restrições: [INFORME O QUE DEVE OU NÃO DEVE APARECER].
O resultado final deve ser entregue obrigatoriamente em dois idiomas: português e inglês, mantendo o mesmo significado e nível de qualidade em ambos. Além disso, gere dois tipos de prompt: Prompt positivo, descrevendo exatamente o que deve acontecer e ser priorizado e Prompt negativo, listando tudo o que deve ser evitado.
O conteúdo deve ser prático, direto, de alta qualidade e totalmente pronto para uso, sem explicações adicionais, sem textos genéricos e sem parecer robótico. Entregue apenas o prompt final, claramente separado por idioma e por tipo (prompt positivo e prompt negativo), escrito de forma organizada e profissional.
Aula Nº 4 - Geração de Imagens de Referencia
PROMPT 1 – Criação da influencer
POSITIVE PROMPT (CONTINUOUS TEXT)
Create an extremely realistic photographic image of an adult sensual woman, approximately 25 years old, designed for a male audience and for Instagram Reels, using a strong UGC (User Generated Content) aesthetic, giving the impression of a real frame captured from a spontaneous video. The model must be blonde, very attractive in a natural and realistic way, with authentic human features and no exaggerated beauty standards. Sensuality should be subtle, elegant, and natural, expressed through posture, gaze, and presence, never through nudity or explicit content. The skin must show true human texture, with visible pores, micro-imperfections, slight natural shine, and subtle tone variations. The face should display a confident, engaging expression, with a natural gaze and a soft smile or casual look, never forced. Blonde hair must look realistic, with visible individual strands, natural volume, and slight movement. Makeup should be light and realistic, enhancing natural facial features without artificial effects. Clothing must be modern, feminine, and discreetly sensual, such as a fitted dress, elegant top, or stylish casual outfit, with realistic fabric and fit. Body proportions must be natural, with a relaxed yet confident posture. Lighting should be soft and natural, like window light or ambient daylight, with realistic shadows and no studio look. The environment should be simple and real, such as a bedroom, living room, neutral wall, or balcony, reinforcing the UGC feel and avoiding any campaign or editorial appearance. The framing must be vertical 9:16, with realistic depth of field, sharp focus, DSLR-quality photography, natural slightly warm colors, and an overall human, authentic, attractive, and trustworthy look, like a perfect still frame from a real Reel.
NEGATIVE PROMPT (CONTINUOUS TEXT)
Completely avoid nudity, explicit content, as well as vulgar or unnatural poses. Do not generate doll-like, CGI, illustrated, anime, cartoon, or 3D appearances. Avoid plastic, overly smooth skin or heavy beauty filters. Do not create overly symmetrical faces, unrealistic beauty, or artificial perfection. Avoid harsh studio lighting, cinematic lighting effects, or unrealistic shadows. Do not use heavy makeup, blur, skin smoothing, or artificial enhancement filters. Avoid transparent, fetishized, or unrealistic clothing. Do not generate distorted body proportions, unnatural poses, or incorrect anatomy. Avoid deformed hands, extra fingers, anatomical errors, or rendering glitches. Do not use artificial or obviously AI-generated backgrounds. Avoid low resolution, lack of sharpness, or a corporate stock photo look.
PROMPT 2 – Close Up de uma imagem
POSITIVE PROMPT (CONTINUOUS TEXT)
Use the uploaded reference image as an absolute and unchangeable base. Generate 10 ultra-realistic close-up images, where each image must contain ONLY ONE unique close-up, filling 100% of the frame, with no splits, collages, grids, multiple panels, or side-by-side comparisons. All images must feature the exact same model from the original image, preserving identical facial identity, apparent age, skin tone and texture, makeup, hair, clothing, and environment, with no aesthetic changes. All images must be created in the same location as the reference photo, with identical and consistent lighting, faithfully matching the original light direction, intensity, shadows, and colors. Each generation must present one exclusive close-up, varying head angle, face tilt, and eye direction, creating 10 distinct and natural poses, like sequential frames from the same real video. The close-ups must include, in separate and unique images: mouth, one eye, the other eye, both eyes together, nose, ear, and additional extremely close facial views, always one main focal area per image, totaling 10 distinct images, all wearing the same outfit and in the same environment. Skin must display true human texture, with visible pores, micro-imperfections, and natural shine. Eyes must show realistic depth, natural reflections, and precise focus. Lips must retain natural texture with no artificial smoothing. Visible hair must show individual strands naturally. Framing must be extremely close yet optically realistic, with no lens distortion, professional DSLR or mirrorless quality, realistic depth of field, and extreme sharpness. Premium UGC style, fully human, spontaneous, and absolutely faithful to the reference image.
NEGATIVE PROMPT (CONTINUOUS TEXT)
Do not generate split images, collages, grids, multi-frame layouts, or side-by-side comparisons. Do not alter facial identity, face shape, age, makeup, hair, clothing, environment, or lighting. Do not repeat identical close-ups without real angle variation. Avoid any CGI, illustration, anime, cartoon, or 3D appearance. Do not apply beauty filters, blur, skin smoothing, or texture reduction. Avoid lens distortion, unrealistic framing, or incorrect proportions. Do not generate artificial eyes, exaggerated shine, or plastic skin. Avoid color shifts, skin tone changes, or visual inconsistencies between images. Do not create anatomical errors, extra parts, or deformations. Avoid low resolution, lack of sharpness, excessive noise, or stock photo appearance.
PROMPT 3 – Imagens em lugares e roupas diferentes
POSITIVE PROMPT (CONTINUOUS TEXT)
Create 10 ultra-realistic photographic images, each image being one single shot (no splits or collages), featuring the same adult female model, approximately 25 years old, blonde, naturally beautiful, with a subtle, elegant sensuality, designed for a male audience and optimized for Instagram Reels (9:16) with a strong UGC (User Generated Content) aesthetic. The model must look fully human and realistic, with authentic facial identity preserved across all images, while changing locations and outfits in every image. Sensuality should be conveyed through posture, gaze, and presence—never explicit, never nude. Each image must take place in a different real-world location, such as: a softly lit bedroom, a modern living room, a balcony at golden hour, an urban café, a minimalist hallway, a sunlit window corner, a rooftop at sunset, a neutral textured wall, a hotel room, and an outdoor city street. Every image must feature a different outfit, all modern, feminine, and discreetly sensual, such as fitted dresses, elegant tops, casual chic outfits, stylish sets, or minimal fashion looks, with realistic fabric behavior and natural fit. The model should display natural micro-poses and variations in body angle, head tilt, shoulder position, and eye direction, making each image feel like a different frame from real UGC videos, never repetitive. Skin must show true human texture, visible pores, micro-imperfections, subtle shine, and realistic color variation. Hair must show individual strands, natural volume, and believable movement. Makeup should be light and realistic, enhancing natural features without artificial effects. Lighting must be natural and environment-appropriate (window light, ambient daylight, soft indoor lighting), with realistic shadows and no studio appearance. Use professional photographic realism with DSLR or mirrorless quality, realistic depth of field, sharp focus, and natural, slightly warm tones. Each image must feel authentic, spontaneous, attractive, and trustworthy—like real content captured casually, not a staged campaign.
NEGATIVE PROMPT (CONTINUOUS TEXT)
Avoid nudity, explicit sexual content, or exaggerated sexualization. Do not generate split images, collages, grids, or multiple frames in a single image. Avoid CGI, illustration, anime, cartoon, or 3D styles. Do not create plastic, overly smooth skin or apply beauty filters, blur, or skin smoothing. Avoid unrealistic beauty standards, overly symmetrical faces, or artificial perfection. Do not use harsh studio lighting, cinematic lighting, or unrealistic shadows. Avoid vulgar poses, forced expressions, or unnatural body positions. Do not repeat outfits or locations. Avoid distorted proportions, lens distortion, anatomical errors, deformed hands, or extra fingers. Do not change the model’s identity between images. Avoid artificial or obviously AI-generated backgrounds. Do not produce low resolution, lack of sharpness, heavy noise, or corporate stock photo appearance.
🔗 Links
https://nordy.ai/comfyui/?flow=66ac86f982cc009a59ed6dce
https://manus.im/invitation/NZENMNIBCWWQQV
