UnityShots

Abstract

Generating a coherent multi-shot video requires structured cross-shot memory: subject appearance, scene context, and speaker identity must persist across cuts. Existing approaches either train end-to-end over fixed-length sequences and cannot scale, generate shot-by-shot with memory banks that grow linearly, or orchestrate pretrained generators under an LLM planner without a multi-shot-aware backbone. We present UnityShots, a memory-driven multi-shot audio-video generation system built on LTX-2.3. The video stream maintains two fixed-size slots — a long-term memory (LTM) slot anchored to the opening shot and a short-term memory (STM) slot holding the immediately preceding tail — both updated by a boundary-conditioned gate at every cut. The audio stream injects a reference speaker token at every shot to preserve vocal timbre without a sliding audio bank. A discrete cut-type prior, learned through AdaLN, becomes an inference-time control knob over transition strength. We release a benchmark of 200 multi-cultural multi-shot sequences spanning six cultural regions and 13 languages. Evaluated across I2V, T2V, and R2V conditioning modes, UnityShots achieves performance comparable to strong open-source and closed-source baselines on cross-shot coherence and audio-video quality.