NiftyNook Cinematic Story Maker · skill pública Open Skill
NiftyNook Cinematic Story Maker
NiftyNook Cinematic Story Maker is an AI-powered storytelling skill designed to transform simple story ideas, plot summaries, or episode concepts into emotionally engaging, cinematic short-form videos. The skill specializes in drama, romance, mystery, suspense, family conflict, relationships, betrayal, secrets, second chances, inspiration, and real-life-inspired stories. It creates stories that feel like professionally directed movie scenes rather than generic AI videos. Every story is structured around a strong opening hook, believable characters, emotional tension, natural dialogue, visual storytelling, escalating conflict, unexpected revelations, and a compelling ending or cliffhanger. The skill supports African, Western, and multicultural characters, including Black, White, and mixed-race characters. It can create stories set in African cities, Western cities, international locations, modern homes, offices, schools, restaurants, hotels, neighborhoods, and other realistic environments. Character consistency is a major priority. Characters should maintain the same facial features, age, hairstyle, skin tone, body type, clothing continuity, personality, and relationships throughout the story unless a deliberate change is required by the plot. Each scene is designed with cinematic direction, including camera angles, camera movement, lighting, environment, facial expressions, body language, character actions, atmosphere, and dialogue. The skill uses techniques such as close-ups, reaction shots, over-the-shoulder shots, establishing shots, slow push-ins, tracking shots, and dramatic pauses to create emotional depth. Stories are optimized for short-form vertical video, particularly Facebook Reels, YouTube Shorts, and TikTok. The first few seconds must immediately create curiosity through a shocking discovery, mysterious situation, emotional confrontation, secret, unexpected visitor, suspicious behavior, or unanswered question. The skill prioritizes realistic human performances, natural dialogue, expressive facial reactions, believable environments, cinematic lighting, realistic skin texture, professional color grading, and high-quality visual continuity. It avoids cartoonish visuals, distorted faces, inconsistent characters, unnatural dialogue, excessive camera movement, unnecessary visual effects, stereotypes, random location changes, and cluttered scenes. Core Story Structure Every story can follow: HOOK → SETUP → CONFLICT → RISING TENSION → CONFRONTATION → REVELATION → CLIFFHANGER/PAYOFF The ending should encourage viewers to continue watching, comment, share, or look for the next episode. Ideal Output The skill can generate: Story title Viral opening hook Character profiles Scene-by-scene storyline Natural dialogue Cinematic visual prompts Camera directions Character expressions and actions Location and lighting instructions Character continuity instructions Scene transitions Suspense and emotional beats Cliffhanger endings Episode captions Thumbnail text Social-media descriptions NiftyNook Brand Identity The visual and storytelling style should feel: Cinematic • Emotional • Sophisticated • Relatable • Suspenseful • Global • Human Brand tagline: Stories That Keep You Watching The overall goal is to create stories about different people, different cultures, and different worlds while focusing on emotions that everyone can understand. Different People. Different Worlds. Same Emotions. Every video should feel like a story worth watching until the very end.
ShortDrama
11 1 8
https://d7qlwi7tl2u27.cloudfront.net/skill-covers/prod/3120234/2026-09-21/u73jw6jh.png
# NiftyNook Cinematic Story Maker ## When to use - The user wants to turn a story idea, plot summary, or episode script into a polished short-form cinematic drama video for vertical platforms (Facebook Reels, YouTube Shorts, TikTok, etc.). - The user wants to revise one stage of an existing NiftyNook story (rewrite the script, regenerate a character, redo a scene video, re-mix the episode). ## Reference files - `references/style-guide.md`: The full NiftyNook story & visual style guide (story style, characters, format, hook, structure, dialogue, cinematic direction, visual style, African/Western settings, emotional acting, continuity, output, branding). Read this before developing any story. ## Preset assets - `brand_logo`: The NiftyNook brand logo image. Use it when branding is requested (see step 10). ## Steps 1. First use `write_todos` to lay out the stages, then update it as you progress. 2. Read `references/style-guide.md` before developing the story. 3. Develop the story: take the user's idea/plot/script and expand it into the 6-scene structure (Hook → Setup → Rising tension → Confrontation → Twist/revelation → Cliffhanger/payoff) with full cinematic direction per scene. Save the complete script as a document with `create_document(content=..., name="<Story Title> — Script", entity_type="scripts")`. 4. Character references: generate one reference image per main character with `generate_image` (model='t2i', orientation='portrait', resolution='720P', name="<Character Name> — Reference", entity_type="characters"). Keep faces, age, skin tone, and hairstyle consistent across all characters. 5. Scene storyboards: for each scene, generate a storyboard image with `generate_image` (model='i2i', input_urls=[the character reference images present in that scene], orientation='portrait', resolution='720P', name="Scene <N> — <title>", entity_type="storyboard_images"). Describe location, time of day, characters, appearance, expression, body language, camera angle, lighting, and atmosphere per the style guide. 6. Scene videos: for each scene, generate a video with `generate_video` (model='i2v', input_urls=[that scene's storyboard image], orientation='portrait', resolution='720P', duration=5–8s, name="Scene <N> — <title>", entity_type="storyboard_videos"). Describe the action, camera movement, and emotional acting per the style guide. 7. Voiceover (optional, if dialogue is wanted): generate per-scene dialogue with `generate_tts` (model='tts.design', text=<dialogue>, voice_instruction=<English voice description>, name="Scene <N> — VO", entity_type="voiceovers"). 8. Music (optional): generate a background score with `generate_music` (prompt=<mood/atmosphere>, name="<Story Title> — Score", entity_type="music"). 9. Mix: stitch all scene videos (and voiceover/music if generated) into the final vertical cut with `mix_clips` (clips=[...], orientation='portrait', name="<Story Title> — Final", entity_type="final_videos"). 10. Branding (only when the user requests it): overlay the `brand_logo` preset asset subtly onto the final cut or thumbnail — never covering faces or important action. Keep it small and unobtrusive. 11. Reply with the story title, hook, character list, scene-by-scene summary, the final video, a short social-media caption, and suggested thumbnail text. Call `set_guidance` to suggest next actions (e.g. tweak a scene, change a character, re-mix, next episode). ## Notes - Default format is 9:16 vertical (orientation='portrait') for Facebook Reels / YouTube Shorts / TikTok. - Maintain strong character consistency: reuse the same character reference images across scenes; do not randomly change clothing, location, time of day, or relationships unless the story requires it. - The first 2–3 seconds must be a hook (shocking discovery, unanswered question, suspicious behavior, emotional confrontation, unexpected message, secret revealed, caught lying, mysterious call, unexpected visitor) — never a slow intro. - Video generation does not