---
title: VideoSearch-R1
emoji: 🔎
colorFrom: blue
colorTo: purple
sdk: gradio
app_file: app.py
pinned: false
license: apache-2.0
short_description: Video retrieval with SQR.
---

<div align="center">

# Welcome to VideoSearch-R1

### Iterative Video Retrieval and Reasoning via Soft Query Refinement

<p>
  <a href="https://github.com/mlvlab/VideoSearch-R1">
    <img src="https://img.shields.io/badge/GitHub-Code-181717?style=for-the-badge&logo=github" alt="GitHub">
  </a>
  <a href="https://mlvlab.github.io/VideoSearch-R1/">
    <img src="https://img.shields.io/badge/Project-Page-2b4f9e?style=for-the-badge" alt="Project Page">
  </a>
  <a href="https://arxiv.org/abs/2607.00446">
    <img src="https://img.shields.io/badge/arXiv-2607.00446-b31b1b?style=for-the-badge" alt="arXiv">
  </a>
  <img src="https://img.shields.io/badge/ECCV-2026-4c6fff?style=for-the-badge" alt="ECCV 2026">
</p>

**VideoSearch-R1** is an agentic framework for video corpus moment retrieval. It unifies inter-video retrieval and intra-video temporal reasoning through a retrieve → verify → refine → ground loop, with **Soft Query Refinement (SQR)** in the continuous query embedding space.

</div>

---

## News

- <img src="https://img.shields.io/badge/2026.06.17-4c6fff?style=flat-square" alt="2026.06.17"> 🎉 VideoSearch-R1 is accepted to **ECCV 2026**.

- <img src="https://img.shields.io/badge/2026.06.20-18a058?style=flat-square" alt="2026.06.20"> Code released.

- <img src="https://img.shields.io/badge/2026.06.20-f59e0b?style=flat-square" alt="2026.06.20"> Trained model checkpoints released.

- <img src="https://img.shields.io/badge/2026.07.01-b31b1b?style=flat-square" alt="2026.07.01"> Paper preprint released on [arXiv](https://arxiv.org/abs/2607.00446).

## Released Resources

| Resource | Status | Link |
|---|---:|---|
| Code | Released | [mlvlab/VideoSearch-R1](https://github.com/mlvlab/VideoSearch-R1) |
| Project page | Released | [mlvlab.github.io/VideoSearch-R1](https://mlvlab.github.io/VideoSearch-R1/) |
| Trained checkpoints | Released | See model repos below |
| Paper preprint | Released | [arXiv:2607.00446](https://arxiv.org/abs/2607.00446) |

## Model Checkpoints

| Dataset | Stage 1 SFT | Stage 2 GRPO |
|---|---|---|
| DiDeMo | [didemo-sft](https://huggingface.co/VideoSearchR1/didemo-sft) | [didemo-grpo](https://huggingface.co/VideoSearchR1/didemo-grpo) |
| Charades-STA | [charades-sft](https://huggingface.co/VideoSearchR1/charades-sft) | [charades-grpo](https://huggingface.co/VideoSearchR1/charades-grpo) |
| ActivityNet Captions | Coming soon | Coming soon |

## Links

- [GitHub repository](https://github.com/mlvlab/VideoSearch-R1)
- [Project page](https://mlvlab.github.io/VideoSearch-R1/)
- [Paper](https://arxiv.org/abs/2607.00446)

## Acknowledgements

VideoSearch-R1 builds on the open-source video-language and reinforcement learning ecosystem, and evaluates on VERIFIED with ActivityNet Captions, DiDeMo, and Charades-STA. We thank the benchmark and dataset creators for making these resources available to the community.