--- name: Animate Story description: "Turn photos and ideas into anime-style visual stories — generate scene-by-scene illustrated narratives" category: Content Creation roles: - cmo - social-manager - designer - x-manager --- # Animate Story Turn photos and ideas into anime-style visual stories — generate scene-by-scene illustrated narratives. You have access to image generation through the OpenLabor API. ## How to Generate Images ``` use images GENERATE '{"prompt":"...","model":"flux-schnell","aspect_ratio":"9:16","num_images":"1"}' ``` ## Styles The same pipeline supports multiple visual styles. Always ask the user which style they want, or default to anime. | Style | Prompt Suffix | |-------|--------------| | Anime/Manga | `anime style, manga aesthetic, vibrant colors, dramatic lighting, studio quality` | | Ghibli | `studio ghibli style, soft watercolor, whimsical, warm lighting, hand-painted feel, nostalgic` | | Comic Book | `comic book style, bold ink outlines, halftone dots, dynamic panels, superhero aesthetic, vivid colors` | | Cyberpunk | `cyberpunk style, neon lights, rain-soaked streets, holographic UI, dark futuristic, blade runner aesthetic` | | Watercolor | `watercolor painting style, soft edges, flowing colors, artistic brushstrokes, delicate, dreamy` | | Pixel Art | `pixel art style, 16-bit retro, clean pixels, limited color palette, nostalgic game aesthetic` | ## Pipeline ### Step 1: Story Input The user provides either: - A **photo** of themselves or a scene (to be anime-fied) - A **text description** of a story they want illustrated - A **topic/theme** ("my morning routine as anime", "startup founder origin story") ### Step 2: Scene Breakdown Break the story into 5-8 scenes. For each scene define: - **Scene number** and **description** (what's happening) - **Type**: "anime" (AI-generated) or "real" (user's original photo) - **Image prompt** for anime scenes - **Caption/text overlay** (short, punchy text for the scene) - **Mood** (epic, chill, romantic, dark, funny) Example breakdown: ``` Scene 1: [anime] Hook — dramatic wide shot establishing the world Scene 2: [real] User's actual photo — the "real me" Scene 3: [anime] The struggle — character facing a challenge Scene 4: [anime] The breakthrough moment Scene 5: [anime] The transformation — before/after contrast Scene 6: [real] User's photo again — "back to reality" Scene 7: [anime] The vision — where they're heading Scene 8: [anime] CTA scene — bold text with call to action ``` ### Step 3: Image Generation For each anime scene, generate the image: ``` use images GENERATE '{"prompt":"[scene description], anime style, manga aesthetic, vibrant colors, dramatic lighting, studio quality, detailed illustration","model":"flux-schnell","aspect_ratio":"9:16","num_images":"1"}' ``` **Style prompt suffix** — always append to every anime prompt: ``` anime style, manga aesthetic, vibrant colors, dramatic lighting, studio quality, detailed illustration, cinematic composition ``` **Aspect ratios by platform:** - TikTok / Reels / Stories: `9:16` (vertical) - Twitter / LinkedIn: `16:9` (horizontal) - Instagram feed: `1:1` (square) **Model selection:** - `flux-schnell` — fast drafts, $0.003/img (~5s) - `flux-dev` — higher quality, $0.025/img (~15s) - `imagen-4` — alternative style, good for text rendering **Generate all scenes in parallel** for speed — don't wait for one to finish before starting the next. ### Step 4: Assemble the Story Present the scenes in order with: 1. Each generated image 2. Scene caption/text overlay 3. Transition notes (if creating video) ### Step 5: Refinement Once the user reviews: - Regenerate specific scenes with adjusted prompts - Try different models for better quality - Adjust style (more dramatic, more subtle, different color palette) - Add or remove scenes ## Story Templates ### "Day in My Life" (anime edition) 1. Wake up scene — cozy bedroom, morning light 2. Real photo — actual morning routine 3. Commute/travel — dramatic city scene 4. Work scene — focused character at desk 5. Real photo — actual workspace 6. Achievement moment — celebration/breakthrough 7. Evening — peaceful sunset scene 8. CTA — "Follow for more" ### "Origin Story" (startup/brand) 1. The problem — dark, frustrated scene 2. The spark — lightbulb moment, dramatic lighting 3. Building — montage of work and progress 4. The struggle — obstacles, rain, darkness 5. Breakthrough — explosion of light and color 6. Today — confident character, bright scene 7. Vision — epic future landscape 8. CTA — brand name, bold typography ### "Product Launch" 1. The old way — dull, frustrating 2. Enter the hero — product reveal, dramatic 3. Feature 1 — action scene showing benefit 4. Feature 2 — another action scene 5. Feature 3 — climactic demonstration 6. Transformation — before/after split 7. Social proof — characters celebrating 8. CTA — "Try it now" ### "Tutorial / How-To" 1. The goal — what you'll learn 2-6. Step-by-step scenes, each illustrating one step 7. The result — completed outcome 8. CTA — "Save this for later" ## Prompt Engineering for Anime ### Character Consistency To maintain consistent characters across scenes: - Describe the character the same way each time: "young woman with short black hair, blue eyes, wearing a white hoodie" - Keep the style descriptors identical across all scenes - Use the same model for all scenes in one story ### Mood Mapping | Mood | Prompt keywords | |------|----------------| | Epic | dramatic lighting, cinematic, golden hour, lens flare, wide angle | | Chill | soft colors, pastel, warm lighting, cozy, lo-fi aesthetic | | Romantic | cherry blossoms, soft pink, dreamy, bokeh, sunset | | Dark | noir, shadows, rain, neon lights, moody, high contrast | | Funny | exaggerated expressions, chibi style, bright colors, comedic | | Hype | action lines, dynamic pose, explosion effects, intense colors | ### Platform Optimization - **TikTok/Reels**: 9:16, bold text overlays, 5-8 scenes max - **Twitter thread**: 16:9, one image per tweet, 7-10 scenes - **Instagram carousel**: 1:1, clean design, 5-10 slides - **LinkedIn**: 16:9, professional tone, 5-7 scenes ## Guidelines 1. Always use 9:16 for vertical/story content unless specified otherwise 2. Generate scenes in parallel for speed 3. Append the anime style suffix to every prompt 4. Start with flux-schnell for drafts, upgrade to flux-dev for finals 5. Keep stories to 5-8 scenes — too many loses engagement 6. Each scene should have a clear purpose in the narrative arc 7. Include a CTA in the final scene