Dream-to-Video AI Cinema

•
267
•

Transforms text-based dream descriptions into high-fidelity cinematic videos using automated prompt engineering and browser-based generation.

Dream-to-Video is a specialized Claude Code Skill that automates the entire pipeline of turning abstract dream logs, diaries, or narratives into surreal, cinematic video content. It utilizes sophisticated prompt engineering to translate text into detailed visual instructions—emphasizing 'dream-logic,' fisheye lens aesthetics, and hardcore realism—while managing browser automation through Playwright to submit tasks to the Jimeng platform. The skill includes automated environment reference mapping, character tagging for consistency, and a custom post-processing engine that applies unique visual effects like 'Elliptic Shatter' to the final output, providing a complete end-to-end creative workflow.

主要功能

01Automated Jimeng platform interaction and video generation via Playwright
02Background worker mode for queuing multiple generation tasks with audio notifications
03Custom post-processing with 'Elliptic Shatter' and cinematic letterboxing
04Surrealist prompt engine with strict cinematic and hardcore realism rules
05267 GitHub stars
06Context-aware environmental reference mapping and character ethnicity tagging

使用场景

01Visualizing personal dream journals or narrative scripts into high-fidelity video
02Prototyping surrealist cinematic concepts using environment-specific reference images
03Automating the creation of atmospheric B-roll for creative media projects