The AI Wedding
Filmmaker Toolkit

Use the attached reference frame or video as the source of truth. Preserve exactly: - the people and their facial identity - wardrobe and accessories - body proportions and pose - camera angle and framing - camera movement - background geometry and location - lens perspective - all important objects in the scene Change only: [DESCRIBE THE LIGHTING / ATMOSPHERE / ENVIRONMENTAL CHANGE] Target look: - Time of day: [TIME OF DAY] - Light direction: [FRONT / SIDE / BACK / OVERHEAD] - Light quality: [SOFT / HARD / DIFFUSED / HAZY] - Color temperature: [WARM / COOL / NEUTRAL] - Mood: [ROMANTIC / CINEMATIC / NATURAL / MOODY] - Contrast: [LOW / MEDIUM / HIGH] The result should look like the same footage captured under different lighting conditions, not a newly generated scene. Keep skin tones natural and realistic. Do not alter faces, wardrobe, composition, architecture, camera movement, or subject position.

Use the source video as the absolute structural reference. Improve only the following: [DESCRIBE WHAT NEEDS IMPROVEMENT] Examples: - lighting - atmosphere - weather - background cleanup - unwanted object removal - reflections - haze - environmental effects Everything else must remain unchanged. Preserve: - people - facial identity - expressions - timing - motion - camera movement - framing - perspective - wardrobe - environment - object placement The goal is to make the original footage look as though the desired change existed practically when it was filmed. Do not reinterpret or redesign the scene.

Generate seamless environmental ambience for: Location: [LOCATION] Environment: [INDOOR / OUTDOOR] Time: [TIME OF DAY] Weather / conditions: [CONDITIONS] Primary background sounds: [LIST] Distance: [NEAR / MEDIUM / DISTANT] Intensity: [VERY SUBTLE / NATURAL / ACTIVE] This audio will be used underneath edited dialogue, so keep it consistent and unobtrusive. Requirements: - no sudden events - no prominent foreground sounds - no intelligible speech - no music - no dramatic changes in volume - consistent texture from beginning to end - seamless enough to loop The result should sound like natural room tone recorded continuously at the location.

Generate a realistic sound effect for: Action: [ACTION] Object: [OBJECT] Surface / material: [MATERIAL] Environment: [ENVIRONMENT] Perspective: [CLOSE / MEDIUM / DISTANT] Microphone feel: [CLEAN / NATURAL LOCATION AUDIO / CLOSE-MIC / CINEMATIC] Intensity: [SUBTLE / NORMAL / STRONG] The sound should feel physically accurate and appropriately scaled. Do not add music, voices, cinematic booms, whooshes, or exaggerated impacts unless explicitly requested. Keep the result usable as natural Foley underneath live-action footage.

Create a new shot that visually belongs in the same film as the attached reference footage. Match the reference in: - lighting direction - time of day - color temperature - contrast - saturation - atmosphere - lens characteristics - depth of field - camera height - perspective - production design - weather - overall cinematic tone Shot needed: [DESCRIBE THE MISSING SHOT] Location: [LOCATION] Subject: [SUBJECT / OBJECT / ENVIRONMENT] Camera: [WIDE / MEDIUM / CLOSE-UP] [LENS OR FOCAL LENGTH IF KNOWN] [CAMERA ANGLE] The final image should feel like it was captured by the same filmmaker, on the same day, with the same visual language as the reference footage. Avoid overly perfect AI-looking lighting, exaggerated depth of field, fantasy details, or anything that would make the shot stand out from the surrounding footage.

Create a photorealistic cinematic establishing shot of: [LOCATION / VENUE / ENVIRONMENT] The purpose of the shot is: [ESTABLISH LOCATION / TRANSITION INTO SCENE / INTRODUCE VENUE / SHOW TIME OF DAY] Scene conditions: - Time of day: [TIME] - Weather: [WEATHER] - Season: [SEASON] - Lighting: [LIGHT DESCRIPTION] Camera setup: - Shot size: wide establishing shot - Camera height: [GROUND / ELEVATED / DRONE] - Lens feel: [ULTRA WIDE / NATURAL / TELEPHOTO] - Composition: [DESCRIPTION] - Movement potential: leave visual depth for [DOLLY / DRONE / ORBIT / PUSH-IN] Visual style: Natural, cinematic, realistic, restrained, and consistent with professional live-action footage. Avoid: - surreal architecture - excessive HDR - artificial symmetry - unrealistic skies - oversaturated colors - random people or objects - obvious AI artifacts

Create a photorealistic cinematic detail shot of: [OBJECT / DETAIL] Context: This shot belongs in a [TYPE OF FILM / EVENT / PROJECT] taking place at [LOCATION / ENVIRONMENT]. Match: - lighting: [DESCRIPTION] - color palette: [DESCRIPTION] - time of day: [TIME] - environment: [DESCRIPTION] Camera: - close-up or macro composition - shallow but believable depth of field - natural lens falloff - realistic materials and textures - tasteful editorial composition The image should feel observational and naturally photographed rather than staged specifically for AI generation. Do not add text, logos, hands, decorative objects, or background elements unless requested.

Animate the attached photograph into a subtle cinematic video shot. Preserve: - facial identity - pose - clothing - composition - background - lighting - architecture - all objects in the original image Camera movement: [VERY SLOW DOLLY IN / DOLLY OUT / LEFT-TO-RIGHT TRACK / ORBIT / PARALLAX] Subject movement: Keep movement minimal and realistic. Allow only subtle natural motion such as: - breathing - blinking - small fabric movement - hair moving gently - environmental motion caused by wind The camera move should be smooth, slow, and physically believable. The result should feel like a shot captured by a real camera, not a photograph being artificially animated. Avoid: - face morphing - changing expressions - new gestures - warped hands - shifting architecture - changing clothing - excessive subject motion

Turn this reference image into a cinematic shot using: Camera movement: [CAMERA MOVE] Speed: [SLOW / VERY SLOW / MEDIUM] Camera behavior: Smooth, stable, physically plausible, and professionally operated. Maintain the exact: - subject identity - environment - lighting - wardrobe - composition - architecture Create realistic parallax based on scene depth. Do not invent new subjects or change objects as the camera moves. Keep the movement restrained enough that the shot can cut naturally with real footage.

Act as an experienced documentary and wedding-film editor. I will give you a transcript from real footage. Your job is not to rewrite the dialogue or invent anything that was not said. Analyze the transcript and propose 3 different editorial structures for a: [TARGET LENGTH] film. The intended emotional tone is: [ROMANTIC / ENERGETIC / INTIMATE / CINEMATIC / FUN / DOCUMENTARY] For each structure provide: 1. Opening What exact dialogue or moment should begin the film? 2. Emotional arc Describe how the story should progress emotionally. 3. Key beats Identify the strongest dialogue moments and put them in a recommended order. 4. Transitions Explain where vows, speeches, ceremony dialogue, interviews, or natural sound could transition between sections. 5. Visual opportunities Suggest what type of B-roll would best support each beat. 6. What to remove Identify redundant, repetitive, or low-value dialogue. 7. Ending Recommend the strongest final emotional beat. Important: - Do not invent quotes. - Preserve the speaker's meaning. - Favor emotional clarity over chronology. - The editor will make the final decisions. TRANSCRIPT: [PASTE TRANSCRIPT]

Act as a professional editor reviewing a rough cut. Below I will provide: 1. the story / transcript 2. the shots or footage I currently have 3. the type of video I am creating Identify any missing visual coverage that would improve the edit. For each missing shot, tell me: - what shot is missing - what story purpose it serves - whether I should ideally use real footage, stock footage, a still image, or generated footage - suggested shot size - suggested camera movement - suggested duration - how important it is: essential / helpful / optional Prioritize shots that: - hide edits - establish location - improve pacing - support dialogue - create emotional breathing room - help transitions Do not suggest generated footage when existing or realistically captured footage would be more authentic. PROJECT: [DESCRIPTION] AVAILABLE FOOTAGE: [LIST] STORY / TRANSCRIPT: [PASTE]

You are a professional cinematographer and AI video prompt designer. Turn my simple request into a detailed production-ready prompt for an image or video generation model. My request: [ROUGH IDEA] Available reference material: [DESCRIBE OR ATTACH] Before writing the final prompt, infer the most important production details while preserving my intent. The final prompt should clearly specify: - subject - action - environment - lighting - time of day - camera angle - shot size - lens feel - camera movement - depth of field - atmosphere - visual continuity requirements - what must remain unchanged - realism requirements - common failure modes to avoid Do not add unnecessary creative elements that I did not ask for. Prioritize: 1. subject consistency 2. physical realism 3. visual continuity 4. clear camera direction 5. concise instructions the generation model can follow Return: 1. FINAL PROMPT 2. PRESERVE / DO NOT CHANGE 3. AVOID