# Complementary Retrieval-Augmented Prompting for Consistent Long-Form Video Generation

arXiv: https://arxiv.org/abs/2609.37407 · HTML: https://mlsanity.com/paper/2609.37407/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Multimodal RAG, one of 15 directions in Retrieval-augmented generation.
- Builds on 19 of its 26 references that are in the map.
- First posted on arXiv 29 Sep 2026.

## Placed in

- Retrieval-augmented generation / [Multimodal RAG](https://mlsanity.com/rag/multimodal-rag/)

## Builds on (most cited first, 19 of 19)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [HPSv3: Towards Wide-Spectrum Human Preference Score](https://mlsanity.com/paper/2508.03789/) (2025)
- [Vidu: a Highly Consistent, Dynamic and Skilled Text-to-Video Generator with Diffusion Models](https://mlsanity.com/paper/2405.04233/) (2024)
- [VideoDirectorGPT: Consistent Multi-scene Video Generation via LLM-Guided Planning](https://mlsanity.com/paper/2309.15091/) (2023)
- [Automated Movie Generation via Multi-Agent CoT Planning](https://mlsanity.com/paper/2503.07314/) (2025)
- [HoloCine: Holistic Generation of Cinematic Multi-Shot Long Video Narratives](https://mlsanity.com/paper/2510.20822/) (2025)
- [StoryMem: Multi-shot Long Video Storytelling with Memory](https://mlsanity.com/paper/2512.19539/) (2025)
- [OneStory: Coherent Multi-Shot Video Generation with Adaptive Memory](https://mlsanity.com/paper/2512.07802/) (2025)
- [LongLive-2.0: An NVFP4 Parallel Infrastructure for Long Video Generation](https://mlsanity.com/paper/2605.18739/) (2026)
- [CineTrans: Learning to Generate Videos with Cinematic Transitions via Masked Diffusion Models](https://mlsanity.com/paper/2508.11484/) (2025)
- [Step-Video-TI2V Technical Report: A State-of-the-Art Text-Driven Image-to-Video Generation Model](https://mlsanity.com/paper/2503.11251/) (2025)
- [ShotStream: Streaming Multi-Shot Video Generation for Interactive Storytelling](https://mlsanity.com/paper/2603.25746/) (2026)
- [VideoMemory: Toward Consistent Video Generation via Memory Integration](https://mlsanity.com/paper/2601.03655/) (2026)
- [LongLive-RAG: A General Retrieval-Augmented Framework for Long Video Generation](https://mlsanity.com/paper/2606.02553/) (2026)
- [EntityBench: Towards Entity-Consistent Long-Range Multi-Shot Video Generation](https://mlsanity.com/paper/2605.15199/) (2026)
- [ReCA: Multi-Shot Long Video Extrapolation via Recursive Context Allocation](https://mlsanity.com/paper/2605.26525/) (2026)
- [Memento: Reconstruct to Remember for Consistent Long Video Generation](https://mlsanity.com/paper/2606.14667/) (2026)
- [GroundShot: Visually Consistent Multi-Shot Long Video Generation via Entity-Grounded Shot Scheduling](https://mlsanity.com/paper/2606.20799/) (2026)
- [EvoState: Closed-Loop Visual State Management for Long-Form Video Generation](https://mlsanity.com/paper/2606.16184/) (2026)

## Abstract (arXiv)

While recent video foundation models excel at generating high-quality short videos, long-form video generation remains a critical challenge, where a major bottleneck lies in conditioning independently generated shots to preserve consistent characters, scenes, and objects throughout a story. Existing training-free approaches typically condition target shots using retrieved historical visuals. However, these references often suffer from severe informational mismatch, either introducing irrelevant contextual redundancy or failing to provide the full combination of required elements for the target shot. To resolve this, we present Complementary Retrieval-Augmented Prompting, an agentic framework that strategically aggregates a compact set of mutually supportive historical references to achieve complete and targeted conditioning for long-form video generation without retraining or modifying the underlying generator. Specifically, our framework explicitly models the visual elements required by each target shot by parsing the narrative script into a text-grounded visual element registry that tracks characters, objects, scenes, and their shot-level states. A VLM-annotated keyframe library further maps these elements to past visual observations. Guided by the required elements, our agent retrieves complementary references that maximize target-element coverage while minimizing historical noise. Finally, the retrieved references, structured element states, and grounding instructions are assembled into a unified prompt for the frozen video generator. This element-aware process provides comprehensive conditioning while remaining fully interpretable. Quantitative and qualitative evaluations on multi-shot story generation demonstrate that our method consistently outperforms recent-frame, memory-based, and entity-level retrieval baselines in cross-shot consistency and text-controllability.