Visual AI engine for image, video, 3D, and audio workflows with GPU scaling and node-based editing.

Hanzo Studio

Hanzo Studio is a visual AI engine for building production image generation, video synthesis, 3D rendering, and audio processing pipelines. It provides a node-based drag-and-drop editor with GPU scaling, distributed workers, multi-tenant isolation, and metered billing.

Dashboard: studio.hanzo.ai Based on: ComfyUI (full node compatibility) Image: ghcr.io/hanzoai/studio:latest

Features

Visual Node Editor — Graph-based canvas for composable AI workflows

GPU on Demand — CPU prototyping, scale to T4/A100/H100 with one API call

Distributed Workers — Route GPU-heavy prompts to dedicated worker machines with automatic queuing

Multi-Tenant Isolation — Per-org storage, compute profiles, and usage tracking via IAM

Metered Billing — Per-prompt billing via Hanzo Commerce, per-second pricing

ComfyUI Compatible — Every custom node, model, and workflow from the ecosystem works out of the box

Workflow Templates — Pre-built workflows for common use cases

Model Zoo — Install models directly from the editor

Architecture

┌─────────────────────────────────────────────────────────────────┐
│                       studio.hanzo.ai                            │
├─────────────────────────────────────────────────────────────────┤
│                                                                  │
│  ┌──────────────┐  ┌───────────────┐  ┌──────────────────────┐  │
│  │  Node Editor  │  │  Coordinator  │  │  Worker Pool         │  │
│  │  (React)      │  │  (Python)     │  │                      │  │
│  │  - Canvas     │  │  - Queue      │  │  worker-0 (T4)      │  │
│  │  - Nodes      │  │  - Route      │  │  worker-1 (A100)    │  │
│  │  - Preview    │  │  - Aggregate  │  │  worker-2 (H100)    │  │
│  └──────────────┘  └───────────────┘  └──────────────────────┘  │
│                                                                  │
│  ┌──────────────────────────────────────────────────────────┐   │
│  │  Model Storage (S3)  │  IAM Auth  │  Commerce Billing    │   │
│  └──────────────────────────────────────────────────────────┘   │
│                                                                  │
└─────────────────────────────────────────────────────────────────┘

Quick Start

Cloud

Go to studio.hanzo.ai

Select a workflow template or start from scratch

Connect nodes, configure parameters, and execute

Self-Hosted

docker pull ghcr.io/hanzoai/studio:latest
docker run -p 8188:8188 ghcr.io/hanzoai/studio:latest \
  python main.py --listen 0.0.0.0 --cpu
# Open http://localhost:8188

GPU

docker run --gpus all -p 8188:8188 ghcr.io/hanzoai/studio:latest \
  python main.py --listen 0.0.0.0 --cuda-device 0

Compute Profiles

Profile	GPU	VRAM	Best For
`cpu`	—	—	Prototyping, lightweight workflows
`gpu-basic`	T4	16 GB	Image generation, standard diffusion
`gpu-pro`	A100	80 GB	Large models, video generation, 3D
`gpu-max`	H100	80 GB	Maximum throughput, batch processing

# Switch compute profile
curl -X PUT https://studio.hanzo.ai/api/compute/config \
  -H "Authorization: Bearer $HANZO_TOKEN" \
  -d '{"active_profile": "gpu-basic"}'

API

Execute Workflow

curl -X POST https://studio.hanzo.ai/api/prompt \
  -H "Authorization: Bearer $HANZO_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": { ... },
    "client_id": "my-client"
  }'

Queue Status

curl https://studio.hanzo.ai/api/queue \
  -H "Authorization: Bearer $HANZO_TOKEN"

Available Models

curl https://studio.hanzo.ai/api/models \
  -H "Authorization: Bearer $HANZO_TOKEN"

Workflow Types

Category	Examples
Image Generation	Stable Diffusion, FLUX, SDXL, ControlNet
Video Synthesis	AnimateDiff, SVD, video-to-video
3D Rendering	TripoSR, instant mesh, point-e
Audio	AudioCraft, Bark, Whisper
Upscaling	ESRGAN, SwinIR, Real-ESRGAN
Inpainting	Segment Anything + diffusion

Environment Variables

Variable	Description
`HANZO_API_KEY`	API key for authentication
`HANZO_IAM_URL`	IAM endpoint (default: `https://hanzo.id`)
`STUDIO_STORAGE_PATH`	Local model storage path
`STUDIO_WORKER_MODE`	`coordinator` or `worker`
`STUDIO_GPU_PROFILE`	Default compute profile
`COMMERCE_BILLING_ENABLED`	Enable per-prompt billing

Complete guide with workflow examples, SDK reference, and advanced configuration

Billing and payment integration for metered usage

Model and asset storage

Authentication and multi-tenant isolation

Hanzo Studio

On this page