--- url: https://superduperai.co/pt/tool/veo3-prompt-generator/ title: "Gerador de prompts para Veo 3" description: "Monte prompts para o Veo 3 na estrutura que o modelo entende melhor: sujeito, ação, cena, câmera, luz, estilo, som ambiente, música, falas e o que evitar. Grátis, no navegador." lang: pt type: tool category: create --- # Gerador de prompts para Veo 3 Um gerador de prompts grátis para Veo 3 e vídeos virais: monta prompts estruturados com câmera, luz, som e fala, roda inteiro no seu navegador e não pede cadastro. O Veo 3 gera vídeo com som, então um bom prompt diz o que aparece, como a câmera se move, o que se ouve e o que deixar de fora. Preencha os campos ou comece de um formato viral; copie o texto ou o JSON. Um gerador de prompts não deixa o modelo mais inteligente: ele impede que você esqueça o que importa. Prompts vagos ("cinematográfico, bonito, 4K") recebem um enquadramento, um movimento e uma luz ao acaso. Este gerador obriga a escolher um de cada, coloca o áudio e a fala onde o Veo 3 espera e deixa os negativos no fim, onde os modelos atuais respeitam melhor. Os formatos prontos são os vídeos curtos que continuam funcionando em 2026: entrevista de rua, ASMR de produto, vlog em primeira pessoa, bicho falante, paródia de plantão de notícias. Troque o sujeito e a fala; mantenha a estrutura. ## Prompt shape (English) {Subject} {action}. {Scene}. {Shot size}, {camera movement}, {lens}. {Lighting}. {Style}. Ambient sound: {ambient}. Music: {music}. Dialogue: {who says what, in quotes}. {Duration} seconds, {aspect}. {Negative: what to avoid}. ## Shot sizes - extreme wide shot - wide shot - medium shot - medium close-up - close-up - extreme close-up - over-the-shoulder shot - POV shot - low-angle shot - high-angle shot - top-down shot ## Camera movements - static, locked-off camera - slow dolly in - slow dolly out - slow push in - tracking shot following the subject - tracking shot alongside the subject - handheld, slight natural shake - slow pan left to right - slow tilt up - slow orbit around the subject - crane shot rising - drone shot, low altitude, slow forward - rack focus from foreground to background - whip pan - selfie-style front camera, arm's length ## Lenses - 24mm wide lens, mild distortion - 35mm lens - 50mm lens - 85mm portrait lens, shallow depth of field - 100mm macro lens - anamorphic lens, oval bokeh - phone camera, wide, deep focus - fisheye lens ## Lighting - golden hour sunlight, long shadows - soft overcast daylight - blue hour, cool ambient light - harsh midday sun - neon practicals, wet reflections - single window light from the left - studio three-point lighting, soft key - candle light, warm flicker - ring light, even front light - moonlight, deep shadows - fluorescent office light, slightly green - volumetric light rays through haze ## Styles - photorealistic, shot on a cinema camera, natural colour - 16mm documentary, light film grain, muted palette - clean commercial look, true-to-life colour, high detail - phone footage, user-generated look, slightly overexposed - VHS 1990s home video, tracking noise, 4:3 feel - 2D anime, cel shading, limited animation - claymation stop motion, 12 fps feel, visible fingerprints - watercolor painting in motion, soft edges, paper texture - pixel art, 8-bit game cutscene - black and white, high contrast, film noir - nature documentary, long lens, slow motion - hyper-real macro food commercial ## Presets (viral formats) ### Entrevista de rua - subject: a young interviewer with a handheld microphone and a stranger in a grey hoodie - action: the interviewer asks a question and the stranger laughs, then answers - scene: a busy city sidewalk at midday, people passing in the background - shot: medium shot - move: handheld, slight natural shake - lens: 35mm lens - light: soft overcast daylight - style: phone footage, user-generated look, slightly overexposed - ambient: street traffic, footsteps, distant chatter - music: no music - dialogue: Interviewer: "What is one thing you would never do again?" Stranger: "Buy a gym membership in January." - negative: no subtitles, no on-screen text, no logos - aspect: 9:16 ### Produto ASMR - subject: a matte black ceramic mug - action: a hand slowly pours steaming coffee into it, steam curls up - scene: a pale oak table by a window, out-of-focus kitchen - shot: extreme close-up - move: slow push in - lens: 100mm macro lens - light: single window light from the left - style: hyper-real macro food commercial - ambient: the pour, a soft clink, room tone - music: no music - dialogue: - negative: no text, no logos, no hands with rings - aspect: 9:16 ### Vlog em primeira pessoa - subject: a woman in her twenties with a backpack, speaking to the camera - action: walks along a harbour, turns the camera to show the boats, then back to her face - scene: a Mediterranean harbour at golden hour - shot: medium close-up - move: selfie-style front camera, arm's length - lens: phone camera, wide, deep focus - light: golden hour sunlight, long shadows - style: phone footage, user-generated look, slightly overexposed - ambient: wind on the microphone, gulls, water - music: no music - dialogue: She says: "Okay, day three. I finally found the place everyone keeps telling me about." - negative: no subtitles, no logos - aspect: 9:16 ### Trailer de cinema - subject: a lone figure in a long coat - action: walks slowly toward the camera through drifting smoke - scene: an abandoned railway station at night - shot: wide shot - move: slow dolly in - lens: anamorphic lens, oval bokeh - light: volumetric light rays through haze - style: photorealistic, shot on a cinema camera, natural colour - ambient: low wind, a distant train horn - music: a single low cello note swelling - dialogue: - negative: no text, no logos, no lens flare overlays - aspect: 16:9 ### Plantão de notícias - subject: a reporter in a navy jacket holding a microphone - action: speaks to the camera, gestures behind her once - scene: outside a modern glass building, people in the background - shot: medium shot - move: static, locked-off camera - lens: 50mm lens - light: soft overcast daylight - style: clean commercial look, true-to-life colour, high detail - ambient: light traffic, wind - music: no music - dialogue: Reporter: "Officials confirmed this morning that the cat has, in fact, been found." - negative: no lower-third graphics, no logos, no on-screen text - aspect: 16:9 ### Bicho falante - subject: a ginger cat sitting upright on a kitchen counter - action: looks at the camera and speaks, mouth moving naturally - scene: a sunny kitchen in the morning - shot: medium close-up - move: static, locked-off camera - lens: 50mm lens - light: single window light from the left - style: photorealistic, shot on a cinema camera, natural colour - ambient: a kettle starting to boil - music: no music - dialogue: The cat says, in a calm deep voice: "I have reviewed your schedule. You are not going to the gym." - negative: no subtitles, no cartoon style - aspect: 9:16 ### Produto em destaque - subject: a pair of white wireless earbuds in an open charging case - action: the case rotates slowly on a turntable, a rim light sweeps across it - scene: a seamless grey studio background - shot: close-up - move: slow orbit around the subject - lens: 100mm macro lens - light: studio three-point lighting, soft key - style: clean commercial look, true-to-life colour, high detail - ambient: silence - music: no music - dialogue: - negative: no text, no logos, no hands, no reflections of people - aspect: 9:16 ### Plano aberto com drone - subject: a fishing village on a fjord - action: one boat leaves a white wake across still water - scene: early morning, ground mist between the houses - shot: extreme wide shot - move: drone shot, low altitude, slow forward - lens: 24mm wide lens, mild distortion - light: blue hour, cool ambient light - style: nature documentary, long lens, slow motion - ambient: distant engine, water, gulls - music: no music - dialogue: - negative: no text, no HDR look, no oversaturation - aspect: 16:9 ### Tutorial com as mãos - subject: two hands and a smartphone on a wooden desk - action: the hands tap through an app, then turn the phone toward the camera - scene: a tidy desk, a plant out of focus - shot: top-down shot - move: static, locked-off camera - lens: 35mm lens - light: ring light, even front light - style: clean commercial look, true-to-life colour, high detail - ambient: soft taps on glass - music: no music - dialogue: Off-camera voice: "Three taps. That is the whole workflow." - negative: no on-screen text, no readable UI text, no logos - aspect: 9:16 ### Cena de anime - subject: a schoolgirl with short dark hair and a red scarf - action: stops on a bridge and looks up as cherry petals drift past - scene: a riverside town in spring, late afternoon - shot: medium shot - move: slow pan left to right - lens: 50mm lens - light: golden hour sunlight, long shadows - style: 2D anime, cel shading, limited animation - ambient: wind, a bicycle bell far away - music: a soft piano motif - dialogue: - negative: no text, no photorealism, no 3D render look - aspect: 16:9 ## FAQ **O que um prompt para o Veo 3 precisa ter?** Um sujeito com detalhes concretos, uma ação, uma cena, um enquadramento e um movimento de câmera, a luz, o estilo e o áudio: som ambiente, música e a fala entre aspas, dizendo quem fala. No fim, o que evitar. **Quanto dura um clipe do Veo 3?** Até 8 segundos. Planeje o vídeo como vários planos de 8 segundos com o mesmo estilo e depois junte tudo na edição. **O Veo 3 gera som e voz?** Sim. Som ambiente, música e fala saem do próprio prompt. Coloque a frase entre aspas e diga quem fala e como. **Posso escrever a fala em português?** Pode. Descreva a cena em inglês, escolha Portuguese em "Idioma da fala" e escreva a frase em português entre aspas; o gerador escreve Dialogue (spoken in Portuguese): "...". Funciona, mas é menos estável do que uma fala em inglês. **Por que meu vídeo saiu com legendas que eu não pedi?** Quando o prompt tem fala, o modelo às vezes desenha legendas. Adicione "no subtitles, no on-screen text" ao negativo. **É grátis? Para onde vai o meu texto?** Grátis e sem cadastro. Tudo roda no seu navegador; nada é enviado a um servidor. Cole o prompt no Veo pelo Gemini, Flow, Vertex AI ou qualquer outra interface. **Serve para outros modelos?** Sim. A mesma estrutura funciona no Kling, Runway e Luma; modelos sem áudio ignoram as linhas de som. ## Related tools - [Video prompt builder](https://superduperai.co/tools/video-prompt-builder/): Fill seven fields, get a text-to-video prompt in the structure Veo, Sora, Kling and Runway read best. - [Script, shot list and prompt list](https://superduperai.co/tools/script-formatter/): Paste a rough script; get timed beats, word count, a shot-list table, a storyboard JSON and one generation prompt per shot. - [Image to prompt](https://superduperai.co/tools/image-to-prompt/): Drop a reference image: palette, brightness, contrast, orientation and composition are read in the browser and written into a prompt skeleton for image-to-video. ## Other languages - en: https://superduperai.co/tool/veo3-prompt-generator/ - es: https://superduperai.co/es/tool/veo3-prompt-generator/ - ru: https://superduperai.co/ru/tool/veo3-prompt-generator/ - tr: https://superduperai.co/tr/tool/veo3-prompt-generator/ - pt: https://superduperai.co/pt/tool/veo3-prompt-generator/