Models and LoRAs
Discover checkpoints, diffusion models, LoRAs, VAEs, samplers, and schedulers from the connected installation.
Choose checkpoints and LoRAs, turn an idea into a production prompt, keep characters consistent, review every result, and let your agents generate images—all while ComfyUI stays on your hardware.

Comfy Studio is free forever—no paid edition, locked features, or account required. If it saved you time and you want to support continued development, you can leave an optional one-time tip. Skipping it changes nothing.
Comfy Studio does not replace ComfyUI. It builds and submits workflows through ComfyUI’s API, then gives everyday generation a focused browser interface.
Discover checkpoints, diffusion models, LoRAs, VAEs, samplers, and schedulers from the connected installation.
Use DeepSeek or another OpenAI-compatible model to select assets and write matched positive and negative prompts.
Save identity anchors, body details, preferred models, LoRAs, triggers, and negative constraints as reusable profiles.
Browse outputs, restore settings, inspect prompts, create variations, favorite assets, and compare results side by side.
Describe what worked or failed after a generation. The writer revises both prompts and can immediately generate the next attempt.
A single endpoint and portable agent skill handle asset selection, prompt writing, generation, waiting, and media delivery.

git clone https://github.com/Pedregoneric/comfy-studio.git cd comfy-studio
COMFY_URL=http://127.0.0.1:8188 npm start
http://127.0.0.1:3030
Private-network beta: Comfy Studio does not yet include application-level authentication. Keep it on localhost, a trusted LAN, or a private tailnet—never the open internet.
Free forever, open source, and ready for humans or agents.