content-engine-dna

Brand DNA files are structured identity files compiled from raw visual assets. They use multimodal analysis to lock visual parameters like color, lighting, and composition, ensuring provenance and brand consistency across multiple AI generation pipelines.

Yes, Gemini multimodal analysis extracts color palettes, lighting, composition, and texture from raw assets. It translates these visual elements into structured parameters to enforce brand and character consistency in downstream generation tools.

You generate optimized prompt fragments by running a compile command on raw assets. The compiler automatically extracts visual parameters and creates tool-specific prompting fragments tailored for platforms like ComfyUI, Nano Banana Pro, Soul Cinema, and Weavy.

Yes, ffmpeg is required to support the content engine pipeline. It helps validate provenance and consistency through a multi-step pipeline including linting and cross-referencing raw source files with compiled identity files.

Style-locking limitations depend on the raw assets provided and the multimodal analysis depth. If source files lack distinct visual parameters or fail automated linting, the compiled identity files may not effectively prevent visual identity drift in downstream tools.