How tillo13/ai_storyboard_video_generator works
Kumori is a script-driven AI storyboard and video generator: users configure story parameters, generate and enrich a chapter-based story, create consistent character and chapter imagery, assemble and post-process videos, then review upload metadata and schedule publication. The README documents the main stages and capabilities; sampled source confirms parts of story generation, image generation, video assembly, and YouTube scheduling. Several stage-to-stage links reflect the documented workflow, while unsampled implementation details are intentionally not inferred.
Main entry points: 9_create_voiceover.py Create voiceover · 0_preview_mosaics_then_upload_to_youtube.py Review and upload · 1_dream_up_a_story.py Generate story · 2_build_out_chapters.py Expand chapters
Components 27
Story creation
- Story settings
GLOBAL_VARIABLES.py - Generate story
1_dream_up_a_story.py - Expand chapters
2_build_out_chapters.py - Summaries and prompts
3_summarize_chapters_add_ai_prompts.py - Story data
Visual generation
- Chapter images
4_create_images_from_ai_prompts.py - Character consistency
4b_unique_character.py - Face identity tools
utilities/faceid_utils.py - Character image generation
utilities/main_character_generator_utils.py - Image enhancement
utilities/enhance_image_via_import.py
Video production
- Create movie
5_create_movie.py - Create mosaic
6_create_mosaic.py - Zoom and pan
7_zoompan_movie.py - Add subtitles
8_add_ffmpeg_subtitles.py - Create voiceover
9_create_voiceover.py - Music sourcing
utilities/rfm_music_utils.py
Publishing
- Review and upload
0_preview_mosaics_then_upload_to_youtube.py - Upload metadata
utilities/youtube_csv_prep_utils.py - Mosaic validation
utilities/mosaic_validator_utils.py - Upload scheduler
utilities/youtube/youtube_scheduler_utils.py - Check scheduled posts
utilities/youtube/youtube_check_publish.py
Other components
- User
- Ollama
- Stable Diffusion
- Google TTS
- FFmpeg
- YouTube
Connections 31
- User → Story settings: sets options
- Story settings → Generate story: configures
- Generate story → Ollama: generates with
- Generate story → Story data: writes story
- Story data → Expand chapters: reads story
- Expand chapters → Ollama: expands with
- Expand chapters → Story data: updates chapters
- Story data → Summaries and prompts: reads chapters
- Summaries and prompts → Ollama: summarizes with
- Summaries and prompts → Story data: writes prompts
- Story data → Chapter images: provides prompts
- Chapter images → Stable Diffusion: generates with
- Story data → Character consistency: provides character data
- Character consistency → Face identity tools: extracts identity
- Character consistency → Character image generation: requests image
- Chapter images → Image enhancement: enhances images
- Story data → Create movie: provides chapter images
- Create movie → FFmpeg: assembles video
- Create movie → Create mosaic: video output
- Create mosaic → Zoom and pan: video input
- Zoom and pan → Add subtitles: video input
- Add subtitles → Create voiceover: video input
- Create voiceover → Google TTS: synthesizes speech
- Create voiceover → Music sourcing: optional music
- Create voiceover → FFmpeg: combines audio
- User → Review and upload: reviews queue
- Review and upload → Mosaic validation: validates mosaics
- Review and upload → Upload metadata: prepares queue
- Review and upload → Upload scheduler: dispatches upload
- Upload scheduler → YouTube: uploads and schedules
- Check scheduled posts → YouTube: checks schedule