Do I need to install anything to use Floyo?

No. Floyo runs entirely in the browser. ComfyUI workflows, models, and nodes are all hosted on Floyo's cloud infrastructure. Users can upload their own ComfyUI workflows and the platform loads all custom nodes and dependencies automatically. There is no local GPU or software requirement.

What does "pay only when you render" mean?

Building, editing, and iterating on workflows is free. GPU charges apply only during active inference — when the workflow is actually generating images or video. There are no charges for idle time, setup, or editing.

Can I use my own custom models and LoRAs?

Yes. Floyo supports uploading custom checkpoints and LoRAs alongside its library of built-in models. The platform is model-agnostic by design — users can mix open-source, proprietary, and custom fine-tuned models within the same workflow.

How does Floyo handle IP and copyright for generated content?

Floyo does not train on user inputs or outputs. All outputs, LoRAs, and workflows belong to the user. Third-party models are accessed through Floyo Partner Nodes with verified no-training commitments from providers.

What GPU hardware does Floyo use?

The standard infrastructure runs on NVIDIA H100 NVL GPUs with 94GB VRAM. Enterprise clients have access to H200 and B200 GPUs for demanding workloads like high-resolution video generation. All compute runs in US-based data centers.

Floyo

Floyo is a browser-based platform for running ComfyUI workflows in the cloud. It targets creative professionals, studios, and enterprises that need GPU-intensive AI generation without local hardware. The platform runs on NVIDIA H100, H200, and B200 GPUs and has been validated by Hollywood studios including Disney, DreamWorks, Paramount, and ILM. Netflix and Amazon have validated its GenAI compliance.

Platform at a Glance

Feature	Details
Core function	Run ComfyUI workflows in-browser
GPU infrastructure	H100 NVL (94GB VRAM), H200, B200
Custom nodes	2,500+ vetted nodes, new ones added daily
Model support	Open-source, proprietary, and custom LoRAs
Model providers	Flux, Kling, Hunyuan, Google, OpenAI, Runway, Luma, Topaz, Recraft, and more
Editing cost	Free (pay only during GPU inference)
Data residency	US-based data centers
IP policy	No training on inputs/outputs; users own all outputs

Four Access Modes

Floyo serves four distinct user types:

Creative professionals. Solo users can build and run any ComfyUI workflow, upload custom checkpoints and LoRAs, and access hundreds of ready-to-run templates. Free to build and edit; pay only when GPUs run inference.

Teams. Shared workspaces with pooled GPU credits, role-based access, and full run history. Designed so that knowledge compounds across the team rather than getting lost in chat tools.

Enterprise. Dedicated GPU pools with full isolation, commercial-use indemnification, custom model and node registries (lock production teams to approved sets while letting R&D test freely), audit logs, and dedicated Slack support with 1-hour target response time.

Developers. Every workflow is exposed as an API — pre-orchestrated, tuned, and ready to call with no infrastructure maintenance required.

IP Protection and Compliance

Floyo positions its IP protection as a core differentiator for studio clients:

No training on inputs or outputs — standard for all users
Verified Private designation — third-party closed-source models run through Floyo Partner Nodes with provider commitments to no training, verified per model in a public Trust Center registry
Commercial Use verified per model — outputs are cleared for commercial production
Full audit trail — every run logged with user, timestamp, and model used
Isolated environments — each run executes in a separate container

This model addresses the copyright and IP ownership concerns that have made studios cautious about AI generation tools.

Free vs Paid

Tier	Price	What’s Included
Free Trial	$0	Run workflows for free, test the platform
Paid Plans	From $12/user/month	More generation time, persistent storage, team workspace
Enterprise	Custom	Dedicated GPU pools, SLA-backed support, studio compliance, onboarding

Building and editing workflows is free on all tiers. Users only pay for GPU time when inference runs. Closed-source models (such as FLUX Pro) require a separate pre-paid API Wallet in addition to the GPU billing.

Floyo Studio

Beyond the self-serve platform, Floyo offers a full-service creative studio. This team has produced work for Toyota, McDonald’s, and Academy Award-winning directors, with alumni from Pixar and ILM. Services include creative direction, production, custom workflow engineering, and team training — bridging the gap between self-serve tools and shot-ready delivery.

Trade-offs to Consider

Floyo requires internet connectivity for all operations — there’s no offline or local mode. The platform is built around ComfyUI’s node-based interface, which has a significant learning curve for users unfamiliar with workflow-style AI tools. Cost predictability depends on understanding how long workflows take to render; complex video generation pipelines can consume GPU time quickly. The credit-based billing for closed-source APIs (separate from GPU billing) adds complexity to cost estimation. Users coming from simpler one-button generation tools may find the workflow construction process overly technical.

Visit Floyo — https://www.floyo.ai/