LLMPopcorn

An LLM-assisted pipeline for generating and evaluating titles, cover prompts, and short-video prompts designed for audience appeal.

LLMPopcorn explores how language models can assist popular micro-video creation. The released pipeline generates titles, cover-image prompts, and short video prompts, then evaluates the resulting content with an offline popularity estimator.

Alongside direct generation, the project includes a prompt-enhancement mode that combines retrieval-augmented generation with step-by-step prompting over the MicroLens dataset.