RunComfy
RunComfy is a cloud-based platform designed to complement and simplify the use of ComfyUI, a powerful and modular stable diffusion GUI used for generating AI ar...
Last verified:
What is RunComfy?
RunComfy is the premier cloud-based ComfyUI platform designed for stable diffusion, empowering AI art creation with high-speed GPUs and efficient workflows without the need for technical setup. It offers a native ComfyUI experience, allowing users to seamlessly transition between local and cloud environments. The platform provides a fully configured ComfyUI Cloud environment that stays in sync with the official ComfyUI repository, meaning everything users are used to locally from custom nodes to downloaded models works exactly the same in the cloud.
Key features include 200+ ready-to-run ComfyUI workflow templates covering video creation, image generation, restoration, upscaling, 3D modeling, and more. RunComfy allows users to download models directly from Civitai, Hugging Face, and Google Drive at speeds 25x faster than local uploads, with models persisted across sessions. The platform offers multiple GPU options ranging from 16GB to 141GB VRAM (T4, A4000, A10G, A5000, A6000, L40S, L40, A100, H100, H200) to handle models of all sizes. Users can install custom nodes freely using ComfyUI-Manager and create reproducible workflow environments that save the entire running setup including workflow JSON, OS, Python environment, custom nodes, and models.
RunComfy is ideal for AI art creators, VFX professionals, educators running classes or workshops, teams collaborating on workflows, and developers needing scalable AI inference. The platform supports building workflows visually in native ComfyUI with zero setup, then deploying any cloud-saved workflow as a Serverless API with autoscaling, zero ops, versioned deployments, and real-time monitoring. Additional offerings include AI Models for instant image/video generation (Seedream 4.0, Flux Kontext Pro, Wan 2.5, Kling 2.5 Turbo, Sora 2), LoRA training on dedicated GPUs (H100, H200, 8xH100), and the RunComfy MCP for AI assistants to directly access Serverless API deployments.
The platform boasts 99.9% uptime, 24/7 support via Discord and email, and is trusted by 1M+ creators including professional courses like ActionVFX's 8-hour ComfyUI course. RunComfy's auto-setup agent handles missing nodes and models automatically when users drop a workflow.json file, skipping dependency setup entirely. Teams can share saved workflows via simple links, giving followers, students, and colleagues instant access to the complete creative environment.
RunComfy pricing
Pricing model: Free
Pay-as-you-go with one shared balance across Models, ComfyUI, and Trainer. Standard CPUSmall machine for uploads/workflow setup: $0.5/hr (free with Pro). GPU rates: T4/A4000 Medium $0.99/hr ($0.79/hr Pro), A10G/A5000 Large $1.75/hr ($1.39/hr Pro), A6000 X-Large $2.50/hr ($1.99/hr Pro), L40S/L40 X-Large Plus $2.99/hr ($2.15/hr Pro, API only), A100 2X-Large $4.99/hr ($3.99/hr Pro), H100 2X-Large Plus $7.49/hr ($5.99/hr Pro), H200 3X-Large $9.59/hr ($7.66/hr Pro). Pro plan: $239.90/year (4 months free) with 20%+ off GPU rates, 20 CPU hrs included, unlimited saved workflows, 200 GB storage, priority support. Training special rates: H100 $4.49/hr ($3.59/hr Pro), H200 $5.75/hr ($4.59/hr Pro), 8xH100 $29.99/hr ($23.99/hr Pro). Model APIs: Wan 2.5 $0.045/second, Kling 2.5 Turbo $0.07/second, Sora 2 $0.10/second, Seedream 4.0 $0.03/image, Flux Kontext Pro $0.04/image. 10 GB storage Pay as You Go (cleared 90 days), 200 GB Pro. Inference deployment rates match GPU machine rates.
RunComfy pros
- No installation required - start creating in seconds
- Native unmodified ComfyUI synced with official releases
- 200+ ready-to-run workflow templates with stunning visuals
- High-speed GPU options from 16GB to 141GB VRAM
- Download models 25x faster from Civitai, Hugging Face, Google Drive
- Models persist across sessions
- ComfyUI-Manager for effortless custom node installation
- Reproducible workflow environments saving entire setup
- Share workflows via simple link with team/followers
- Serverless API deployment with autoscaling and zero ops
- Auto-setup agent handles missing nodes and models automatically
- 99.9% uptime guarantee
- 24/7 support via Discord and email
- Pay-as-you-go billing with auto-release GPUs, no idle charges
- Pro plan offers 20%+ off all GPU rates
- One shared balance across Models, ComfyUI, and Trainer
- Perfect for classes, workshops, tutorials with consistent environment
- Support for full stack: instant generation demos plus LoRA training
- Versioned deployments with real-time monitoring and run history
- Trusted by 1M+ creators including VFX professionals
RunComfy cons
- Pay-as-you-go pricing can be expensive for heavy usage
- Pro plan requires $239.90/year subscription for discounts
- Persistent storage cleared after 90 days on Pay as You Go
- Only 10 GB storage on Pay as You Go vs 200 GB on Pro
- Only up to 10 saved workflow environments on Pay as You Go
- Standard support limited to Discord and email, no priority
- Some GPU machines API-only (L40S/L40 X-Large Plus)
- Training runs require dedicated GPU booking
Frequently asked questions about RunComfy
What is ComfyUI?
ComfyUI is the most powerful and versatile graphical user interface (GUI) for stable diffusion, offering a comprehensive API and backend architecture. It supports SD1.x, SD2, SDXL, SD3, SD3.5, Flux.1, AnimateDiff, controlnet, Stable Video Diffusion, and many others. Its intuitive node-based graph system empowers users to design and experiment with intricate image and video creation workflows effortlessly, providing precise control over every aspect without prior coding experience.
How can I get started with ComfyUI?
Installing ComfyUI locally can be complex and requires a powerful GPU. RunComfy provides a fully configured ComfyUI cloud environment ready to use right away, letting you focus on building and experimenting with creative workflows. You can get started faster with 200+ ComfyUI workflow templates. For help or community connection, reach out on Discord.
Is there a ComfyUI API service?
Yes. On RunComfy you can turn any cloud-saved ComfyUI workflow into a serverless, autoscaling API in minutes with no DevOps required. You get a stable HTTPS endpoint for programmatic calls, parameterize workflow inputs, receive reproducible outputs with versioning, real-time monitoring, run history, and cost-efficient pay-as-you-go GPU billing. It scales up automatically under load and pays only while jobs run.
Can I upload and download my own models or LoRAs to RunComfy? Will they save after shutting down?
Absolutely! You can directly download models from Civitai, Hugging Face, and Google Drive to RunComfy at speeds 25 times faster than local uploads. You can also upload your own models from your local device. All models are securely saved across sessions, so you won't lose work when closing the machine.
Can I install and manage nodes within RunComfy ComfyUI cloud?
Absolutely! RunComfy provides a pre-installed ComfyUI Manager for easy node installation and management. You can have unlimited ComfyUI environments with different node setups - imagine many cloud machines for playing with different configurations.
Can I upload my images and videos in RunComfy ComfyUI cloud?
Yes, you can upload images and videos (including folder structures) into the input folder of RunComfy ComfyUI using the file browser.
Will generated, uploaded, or downloaded data persist across machine sessions?
Yes! With RunComfy's Cloud Save feature, you save your entire workflow including custom nodes, models, and complete ComfyUI running environment. Next time you open your workflow in RunComfy, everything will be exactly as you left it.
Is there any technical support available?
Yes! Technical support is offered through Discord channel and email at [email protected]. The team assists with any questions or issues while using ComfyUI.
Why should I use RunComfy for my AI class/workshop/tutorials?
RunComfy is great for classes, workshops, and tutorials because students start ComfyUI instantly in the cloud with no local installs, and you get a consistent environment for everyone. It supports the full stack: use Models for instant image/video generation demos, and Trainer for LoRA/fine-tuning on dedicated GPUs. Billing is flexible with one shared balance, pay-as-you-go, plus optional Pro discounts.
Do you have education, team, or enterprise plans?
Yes, RunComfy has collaborated extensively with educational institutions, government-led AI training programs, and enterprises to provide comprehensive AI Art online content and learning environments. Contact them for more information on these plans.