Last updated 08-14-2026
Category:
Reviews:
Join thousands of AI enthusiasts in the World of AI!
Diffus (formerly GRAVITI Diffus)
Diffus lets you generate and edit images in the browser with Stable Diffusion, SDXL, SD3, and Flux models. Pick from 70,000+ checkpoints, LoRAs, and embeddings, then run cloud generations without buying your own GPU. The workspace bundles ControlNet, inpainting, ADetailer face fixes, custom model uploads, ComfyUI workflows, and standalone tools for Flux and video output.
Where Midjourney or DALL·E lock you into one model stack, Diffus behaves more like a hosted Stable Diffusion lab. You can browse community models, upload private ones on higher tiers, and dial in sampling steps, inference time, and concurrent jobs per plan. That depth trades simplicity for control, which suits artists who already speak prompt syntax and extension names.
Designers, illustrators, and marketing teams use Diffus for concept art, product mockups, face-corrected portraits, and short video clips. Hobbyists on the Teaser plan get daily credits to experiment, while Pro subscribers get finetuning, 40GB private storage, and ComfyUI included for repeatable pipelines.
Browse and run 70,000+ public checkpoints, LoRAs, and embeddings from one browser workspace
ControlNet, inpainting, ADetailer, and Couple extensions for composition and face refinement
Plus and Pro plans unlock SD3, Flux, video generation, and up to 36,000 monthly credits
Upload private models on Plus (25GB storage) and Pro (50GB storage) tiers
ComfyUI workflows included on Pro; available as an addon on Basic and Plus
Pro plan supports finetuning and advanced video models with sound
70,000+ community checkpoints, LoRAs, and embeddings without local GPU setup.
Deep Stable Diffusion controls: ControlNet, inpainting, ADetailer, and custom model uploads.
Flux, SD3, video generation, and ComfyUI on higher tiers for production workflows.
Clear credit limits and storage caps spelled out per plan on the pricing page.
No free tier; the entry Teaser plan still costs $1 per month with public-only outputs.
Teaser and Basic plans lack video generation, Flux, and private model uploads.
Support is Discord-only on paid tiers, with no support on the cheapest plan.
How much does Diffus cost?
Diffus offers four paid plans. Teaser is $1 per month with 30 daily credits. Basic is $12 per month on yearly billing with 3,000 monthly credits. Plus is $24 per month with 12,000 credits and Flux access. Pro is $48 per month with 36,000 credits, ComfyUI included, and finetuning.
Does Diffus need a local GPU?
No. Diffus runs image and video generation in the cloud through your browser. You do not need to install CUDA drivers or maintain local hardware. Plans set limits on concurrent jobs, sampling steps, and inference time instead.
What models does Diffus support?
Diffus supports Stable Diffusion 1.5 and SDXL on every plan. Basic adds SD3. Plus and Pro add Flux checkpoints, LoRAs, and embeddings. You can also upload private models on Plus and Pro for niche styles.
Can Diffus generate videos?
Yes. Diffus Plus includes text-to-video and image-to-video generation with models like Kling, Veo, and Wan. Pro unlocks advanced video features with sound and extras. Lower tiers focus on still images only.
Does Diffus include ComfyUI?
Diffus Pro includes ComfyUI for building custom node workflows at no extra charge. Basic and Plus subscribers can add ComfyUI as a paid addon. Teaser does not include ComfyUI access.
How do I get support on Diffus?
Diffus routes support through Discord. Basic gets limited Discord help, Plus gets standard Discord support, and Pro gets priority Discord support. Teaser has no support channel. You can also email [email protected].