Comfy Cloud Native Mobile App Launches via Public API
ComfyUI developed a native mobile application that connects directly to Comfy Cloud via a three-call public API. This architecture enables mobile users to trigger intensive AI video and image generations without local hardware constraints.
ComfyUI, the node-based interface for Stable Diffusion and generative media, launched a native mobile application framework for Comfy Cloud. The implementation demonstrates how the platform's public API can bridge the gap between heavy desktop workflows and mobile devices. By offloading the rendering and processing to cloud-based GPUs, the app allows creators to execute complex workflows—including image generation and video synthesis—on standard iOS and Android hardware.
What's new
ComfyUI's mobile implementation utilizes a streamlined API structure consisting of three primary calls: prompt submission, status polling, and output retrieval. As of February 2025, this architecture supports the full range of Comfy Cloud features, including custom nodes and proprietary workflows. The mobile app is built using React Native, ensuring cross-platform compatibility while maintaining the performance of a native UI.
Unlike previous web-wrapper attempts, this native ComfyUI mobile app handles authentication through Comfy Cloud and provides a dedicated interface for adjusting workflow parameters. The system is designed to handle asynchronous tasks, meaning a user can initiate a high-resolution video render on their phone, close the app, and receive a notification once the file is ready for download. This release serves as a reference implementation for developers looking to build specialized mobile tools on top of the ComfyUI ecosystem.
How it fits your workflow
ComfyUI on mobile changes the utility of the platform for field-based creators and production supervisors. Instead of being tethered to a workstation or a remote desktop session, editors can now trigger pre-configured upscaling or rotoscoping workflows directly from a mobile device. This setup functions as a portable control surface for a creator's existing ComfyUI library, making it an alternative to the mobile web interfaces of Midjourney or Leonardo.ai.
For professional VFX artists and animators, the ComfyUI mobile app acts as a remote rendering queue. While tools like Krea or Runway offer mobile-friendly web versions, the ComfyUI native app approach allows for deeper customization of the underlying node graph. A director on set could, for example, take a reference photo with their phone and immediately run it through a specific ComfyUI 'Style Transfer' workflow to preview a look before the dailies are processed. This integration bridges the gap between high-end desktop AI video generation and the immediacy of mobile capture.
What it costs / how to try it
The mobile application is currently available as a reference project for Comfy Cloud subscribers. Users can access the source code and implementation details via the official ComfyUI blog and GitHub repository to deploy their own instances or use the provided build.
Read the original announcement on ComfyUI ↗