ComfyUI_RunpodNodes

ComfyUI RunPod Nodes

A comprehensive set of ComfyUI custom nodes for interacting with RunPod's serverless API, including both public endpoints (Flux, Qwen, Seedream, WAN, SORA, etc.) and custom serverless workers.

Features

• 27 Public Endpoint Nodes - Ready-to-use nodes for RunPod's public AI models
• Custom Serverless Nodes - Connect to your own RunPod workers
• Image/Video/Audio Support - Generate images, videos, and audio
• Async Job Handling - Automatic polling with status updates

Installation

1. Clone this repository into your ComfyUI custom nodes folder:

   cd ComfyUI/custom_nodes
   git clone https://github.com/your-repo/comfyui-runpod.git
   

2. Restart ComfyUI

Requirements

• RunPod API key (get one at runpod.io)
• ComfyUI

Node Categories

RunPod/Image (10 nodes)

| Node | Description | |------|-------------| | Flux Dev | High-quality image generation with exceptional prompt adherence | | Flux Schnell | Fast image generation, ideal for prototyping | | Flux Kontext Dev | Image editing based on text instructions | | Qwen Image | Image generation with advanced text rendering | | Qwen Image LoRA | Image generation with LoRA support | | Qwen Image Edit | Image editing with text rendering capabilities | | Seedream 3.0 | Bilingual image generation (Chinese-English) | | Seedream 4.0 T2I | New-generation text-to-image | | Seedream 4.0 Edit | Advanced image editing (supports multiple input images) | | Nano Banana Edit | Combine multiple images into a scene (supports 2 input images) |

RunPod/Video (11 nodes)

| Node | Description | |------|-------------| | InfiniteTalk | Audio-driven talking/singing video generation | | Kling v2.1 I2V Pro | Professional image-to-video generation | | Seedance 1.0 Pro | High-performance video with multi-shot storytelling | | SORA 2 I2V | OpenAI's Sora 2 video generation | | SORA 2 Pro I2V | OpenAI's Sora 2 Pro (professional grade) | | WAN 2.5 | Image-to-video generation | | WAN 2.2 I2V 720p LoRA | Video generation with LoRA support | | WAN 2.2 I2V 720p | Image-to-video at 720p | | WAN 2.2 T2V 720p | Text-to-video at 720p | | WAN 2.1 I2V 720p | Image-to-video at 720p | | WAN 2.1 T2V 720p | Text-to-video at 720p |

RunPod/Audio (2 nodes)

| Node | Description | |------|-------------| | Whisper V3 Large | Speech-to-text transcription | | Minimax Speech 02 HD | Text-to-speech with emotion control |

RunPod/Custom Serverless (3 nodes)

| Node | Description | |------|-------------| | RunPod SDXL (Custom Endpoint) | Connect to your own SDXL worker | | RunPod SD (Custom Endpoint) | Connect to your own SD 1.5 worker | | RunPod ComfyUI Workflow Runner | Run exported ComfyUI workflows on RunPod workers |

RunPod/Utilities (1 node)

| Node | Description | |------|-------------| | RunPod Variables Builder | Build JSON variable objects for workflow runners |

Usage

Public Endpoints

1. Add any public endpoint node (e.g., "Flux Dev") to your workflow
2. Enter your RunPod API key
3. Configure the parameters (prompt, size, etc.)
4. Connect to a Preview Image or Save Image node

Example - Flux Dev:

[Flux Dev] → [Preview Image]
     ↑
 api_key: "your-runpod-api-key"
 prompt: "A serene mountain landscape at sunset"
 width: 1024
 height: 1024

Image Editing Nodes

Nodes like Flux Kontext Dev, Qwen Image Edit, Seedream 4.0 Edit, and Nano Banana Edit accept optional image inputs:

• Single image input: Connect an IMAGE to input_image
• Multiple images (Nano Banana, Seedream Edit): Connect to input_image_1 and input_image_2

Video Nodes

Video nodes return a URL string. Use a "Load Video from URL" node or download the video:

[WAN 2.5] → [Save Text] or custom video loader
     ↑
 video_url output → "https://..."

⚠️ Video/audio URLs expire after 7 days. Download immediately if you need to keep them.

Custom Serverless Workers

For your own RunPod workers:

1. Deploy a serverless endpoint on RunPod
2. Use the appropriate node (SDXL, SD, or ComfyUI Workflow Runner)
3. Enter your endpoint ID and API key

ComfyUI Workflow Runner:

1. Export your workflow: Workflow > Export (API)
2. Paste the JSON into workflow_json
3. Optionally connect an input image
4. The node sends the workflow to your worker-comfyui endpoint

API Key

All nodes require a RunPod API key. You can:

• Enter it directly in each node
• Create a reusable "Primitive" node with your API key

Get your API key at: https://www.runpod.io/console/user/settings

Pricing

Public endpoints are billed per use. See RunPod's pricing for current rates:

• Image generation: ~$0.02-0.03 per megapixel
• Video generation: ~$0.30-1.20 per video
• Audio: ~$0.05 per 1000 characters

Troubleshooting

Node appears to "hang"

• Public endpoints can take 30-120 seconds depending on server load
• Check the ComfyUI console for status updates:

  [RunPod] Submitting job to black-forest-labs-flux-1-dev...
  [RunPod] Job submitted: abc123
  [RunPod] Status: IN_QUEUE (elapsed: 0s)
  [RunPod] Status: IN_PROGRESS (elapsed: 15s)
  [RunPod] Job completed successfully
  

"Job failed" errors

• Verify your API key is valid
• Check you have sufficient credits
• Some endpoints require specific image formats or sizes

Large images fail

• RunPod has request size limits (10MB for async, 20MB for sync)
• Use smaller images or URL inputs instead of base64

Links

• RunPod Documentation
• Public Endpoint Reference
• worker-comfyui

License

MIT