Fluid Imagine

Fluid Imagine turns your phone into a remote control for ComfyUI. Type a prompt, tap generate, and watch the image render on your own GPU. Nothing goes to the cloud.

PROJECTS, NOT JUST PROMPTS Each idea lives in its own project with its prompt, settings, and every image you've made from it. Swipe through results full screen, compare variations side by side in grouped view, and pick up where you left off.

READY-MADE WORKFLOWS Start with built-in SDXL, Z-Image Turbo, and Krea 2 Turbo workflows. Fluid Imagine matches them to the models already on your server, so there is no graph editing and no JSON.

PROMPT REFINER Write a few words and let a text model on your server expand them into a detailed prompt.

FINE CONTROL WHEN YOU WANT IT Adjust the sampler, steps, CFG, seed, and image size in the details editor, or leave the defaults alone and just generate.

FINDS YOUR SERVER Scan your Wi-Fi network for ComfyUI, or enter an address directly. Save several servers and switch between them.

PRIVATE BY DESIGN No account, no analytics, no ads. Your prompts and images go only between your phone and the server you choose.

TRY IT FIRST No server yet? Demo mode lets you explore the app with bundled sample images.

UPGRADING FROM FLUID DIFFUSION Your saved templates come across automatically as projects.

REQUIREMENTS • A computer running ComfyUI, started with --listen so your phone can reach it • Compatible models installed on that server (SDXL checkpoint, Z-Image Turbo, or Krea 2 Turbo) • Android 10 or newer