Multimodal Agent MemorySubmit a paper

LongVideoAgent

LongVideoAgent: Multi-Agent Reasoning with Long Videos

Runtao Liu, Ziyi Liu, Jiaqi Tang, Yue Ma, Renjie Pi, Jipeng Zhang, et al. · 2025-12-23

Read the paperPDF

Management
T1 Static Memory
Memory is built once and reused without change during deployment.
Primary modality
Video & streaming
Also: text
Representation
Hybrid
Distinct operative representation paths are deliberately linked or jointly consumed by downstream reasoning, generation, or control.

How experience becomes memory

Each path is one memory component, traced from what is stored to how it is found again.

  1. question-relevant video segments

    1. Represented asRetained perceptual source evidenceSource- and modality-specific
    2. Stored inFlat sequential records
    3. Retrieved bySimilarity search, Metadata / spatiotemporal retrieval, Planner / tool-assisted retrieval
    4. Source evidenceSource directly available
    Grounding agent outputs (segment localizations) and vision agent outputs (textual observations) accumulated by the master agent. iterative retrieval of sparse, question-relevant evidence via the grounding agent; subtitles complemented with visual detail from the vision agent.Full paper: Method · PDF p. 3 · Figure 1 / method overview

  2. targeted textual observations

    1. Represented asPerception-to-text records & summariesCross-represented explicit
    2. Stored inFlat sequential records
    3. Retrieved bySimilarity search, Metadata / spatiotemporal retrieval, Planner / tool-assisted retrieval
    4. Source evidenceSource not retained
    Grounding agent outputs (segment localizations) and vision agent outputs (textual observations) accumulated by the master agent. iterative retrieval of sparse, question-relevant evidence via the grounding agent; subtitles complemented with visual detail from the vision agent.Full paper: Method · PDF p. 3 · Figure 1 / method overview

Related systems