[NeurIPS' 2025] JarvisArt: Liberating Human Artistic Creativity via an Intelligent Photo Retouching Agent

Yunlong Lin^1*, Zixu Lin^1*, Kunjie Lin^1*, Jinbin Bai⁵, Panwang Pan⁴, Chenxin Li³, Haoyu Chen², Zhongdao Wang⁶, Xinghao Ding^1†, Wenbo Li^3♣, Shuicheng Yan^5†

¹Xiamen University, ²The Hong Kong University of Science and Technology (Guangzhou), ³ The Chinese University of Hong Kong, ⁴Bytedance, ⁵National University of Singapore, ⁶Tsinghua University

💡 Our new work that may interest you ✨.

JarvisEvo: Towards a Self-Evolving Photo Editing Agent with Synergistic Editor-Evaluator Optimization
Yunlong Lin, Lingqing Wang, Zixu Lin and Kunjie Lin, etc.

---

📮 Updates

• [2025.12.8] The evaluation set MMArt-Bench is now released! Data construction scripts are now released! Check out Data Scripts
• [2025.12.7] Training (SFT & GRPO-R) and Evaluation scripts are now released! Check out Training Guide and Evaluation.
• [2025.10.7] Local client now supports Agent-to-Lightroom Protocol ! See our Agent-to-Lightroom Protocol documentation for seamless AI agent integration with Adobe Lightroom.
• [2025.10.1] MMArt-PPR10k is now live on Hugging Face Datasets! Built upon @PPR10K, this open-source dataset contains diverse user instructions, alongside Lightroom Lua/XMP files and corresponding original and edited images. It's released under the Apache 2.0 license.
• [2025.9.18] Congratulations! JarvisArt is accepted to NeurIPS 2025.
• [2025.7.14] Thanks to @pydemo for writing a helpful tutorial: Automate Your Lightroom Preset Creation with AI.
• [2025.7.12] Inference code is now available! Check out our Inference documentation.
• [2025.7.9] We're grateful to @AK for featuring JarvisArt on Twitter!
• [2025.7.4] See our Chinese blog to get more details about JarvisArt! 中文解读｜修图界ChatGPT诞生！JarvisArt：解放人类艺术创造力——用自然语言指挥200+专业工具.
• [2025.7.3] Hugging Face online demo is now available: Try it here: JarvisArt-Preview.
• [2025.6.28] Gradio demo and model weights are now available! Check out our Gradio Demo and Model Weights.
• [2025.6.20] Paper is now available on arXiv.
• [2025.6.16] Project page is live.

---

🧭 Navigation

• Overview
• Demo Videos
• Checklist
• Getting Started
  • Gradio Demo
  • Batch Inference
  • Agent-to-Lightroom Protocol
  • Training Guide
  • Data Scripts
  • Evaluation
• Discussion Group
• Citation

---

📝 Overview

JarvisArt workflow and results showcase

JarvisArt is a multi-modal large language model (MLLM)-driven agent for intelligent photo retouching. It is designed to liberate human creativity by understanding user intent, mimicking the reasoning of professional artists, and coordinating over 200 tools in Adobe Lightroom. JarvisArt utilizes a novel two-stage training framework, starting with Chain-of-Thought supervised fine-tuning for foundational reasoning, followed by Group Relative Policy Optimization for Retouching (GRPO-R) to enhance its decision-making and tool proficiency. Supported by the newly created MMArt dataset (55K samples) and MMArt-Bench, JarvisArt demonstrates superior performance, outperforming GPT-4o with a 60% improvement in pixel-level metrics for content fidelity while maintaining comparable instruction-following capabilities.

---

🎬 Demo Videos

Global Retouching Case

Local Retouching Case

JarvisArt supports multi-granularity retouching goals, ranging from scene-level adjustments to region-specific refinements. Users can perform intuitive, free-form edits through natural inputs such as text prompts and bounding boxes

---

💻 Getting Started

For gradio demo running, please follow:

• Gradio Demo

For batch inference, please follow the instructions below:

• Batch Inference

For Agent-to-Lightroom Protocol, please follow:

• Agent-to-Lightroom Protocol

For training (SFT & GRPO-R), please follow:

• Training Guide

For data construction pipeline (image pairs, instructions, CoT generation & format conversion), please follow:

• Data Scripts

For evaluation, please follow:

• Evaluation

---

🎪 Checklist

• Create repo and project page
• Release preview Inference code and gradio demo
• Release huggingface online demo
• Release preview model weight
• Release Agent-to-Lightroom Protocol
• Release MMArt-PPR10K dataset with open license
• Release SFT training code
• Release GRPO-R training code
• Release evaluation code
• Release MMArt-Bench
• Release data construction scripts

---

🔍 Jarvis Family

JarvisIR: An intelligent image restoration agent for diverse and complex degradations in real-world scenarios.

We are excited to expand the Jarvis family with more intelligent agents in the future. Stay tuned for upcoming releases!

🙏 Acknowledgements

We would like to express our gratitude to LLaMA-Factory, gradio_image_annotator and VLM-R1 for their valuable open-source contributions which have provided important technical references for our work.

🌤️ Discussion Group

If you have any questions during the trial, running or deployment, feel free to join our WeChat group discussion! If you have any ideas or suggestions for the project, you are also welcome to join our WeChat group discussion!

Scan QR code to join WeChat group discussion

📧 Contact

For any questions or inquiries, please reach out to us:

• Yunlong Lin: [email protected]
• Zixu Lin: [email protected]
• Kunjie Lin: [email protected]
• Panwang Pan: [email protected]
• Chenxin Li: [email protected]

---

📚 Citation

If you find JarvisArt useful in your research, please consider citing:

@article{jarvisart2025,
title={JarvisArt: Liberating Human Artistic Creativity via an Intelligent Photo Retouching Agent}, 
      author={Yunlong Lin and Zixu Lin and Kunjie Lin and Jinbin Bai and Panwang Pan and Chenxin Li and Haoyu Chen and Zhongdao Wang and Xinghao Ding and Wenbo Li and Shuicheng Yan},
      year={2025},
      journal={arXiv preprint arXiv:2506.17612}
}

---

📜 License

JarvisArt is released under the Apache License 2.0, but commercial use is explicitly prohibited. While the Apache 2.0 license typically allows free use, modification, and distribution of code, for this project we specifically declare: commercial use of JarvisArt and its related code, models, and datasets is forbidden.

Our MMArt-PPR10k dataset is also prohibited from commercial use. Any commercial application requires explicit written permission from the authors.

For commercial cooperation or commercial licensing, please contact the project authors.