How manasi-rathod/storytrails works
StoryTrails turns a story into a scene plan and per-scene narration, with a separate editor that combines available scene media into a movie. A standalone generation script demonstrates SDXL image creation followed by Stable Video Diffusion, but the sampled source does not connect it to the main workflow. The README describes a Streamlit interface; sampled main.py is a script, so that interface is not shown as an implemented component.
Main entry points: final_editor.py Video editor · main.py Workflow script · project_be.py Visual generation demo · director.py Scene planner
Components 15
Story planning
- Workflow script
main.py - Scene planner
director.py
Media generation
- Voice synthesis
narrator.py - Visual generation demo
project_be.py
Movie assembly
- Video editor
final_editor.py
Project artifacts
- Shot list JSON
master_shot_list.json - Voice sample
my_voice.wav - Scene audio
project_assets - Scene video
project_assets - Prepared scenes
temp_scenes - Final movie
FINAL_STORY_MOVIE.mp4
Other components
- Story author
- Ollama LLM
- SDXL image model
- SVD video model
Connections 15
- Story author → Workflow script: provides story
README.md - Workflow script → Scene planner: requests plan
main.py - Scene planner → Ollama LLM: queries model
director.py - Workflow script → Shot list JSON: writes plan
main.py - Workflow script → Voice synthesis: requests narration
main.py - Workflow script → Voice sample: supplies sample
main.py - Voice synthesis → Scene audio: writes audio
- Visual generation demo → SDXL image model: generates image
project_be.py - Visual generation demo → SVD video model: generates video
project_be.py - SDXL image model → SVD video model: provides image
- Video editor → Shot list JSON: reads order
final_editor.py - Video editor → Scene audio: reads audio
final_editor.py - Video editor → Scene video: reads video
final_editor.py - Video editor → Prepared scenes: writes merged scenes
final_editor.py - Video editor → Final movie: writes movie
final_editor.py