ComfyUI_API工具

Install Python dependencies

Install the packages from requirements.txt with the same Python environment that runs ComfyUI:

pip install -r requirements.txt

Dependencies used by this project:

• tenacity
• openai
• pillow
• requests
• numpy
• tiktoken
• urlextract

After installation, restart ComfyUI so the custom nodes are discovered correctly.

Installation

Option 1: Install with ComfyUI Manager

Search for billbum in ComfyUI Manager and install the node pack.

Option 2: Install from Git URL

If you use Install via Git URL in ComfyUI Manager, paste this repository URL:

https://github.com/AhBumm/ComfyUI_BillBum_APIset_Nodes.git

If you prefer a manual install, clone it into your ComfyUI custom_nodes directory:

git clone https://github.com/AhBumm/ComfyUI_BillBum_APIset_Nodes.git

Quick Start

1. Install the node pack and Python dependencies.
2. Restart ComfyUI.
3. Open the example workflow file minimal_api_group.json.
4. Replace the placeholder API keys and base URLs with your own provider settings.
5. Choose a model that is valid for the provider and endpoint you are using.
6. Run one of the minimal workflow groups to confirm your setup.

The bundled example workflow includes these starter groups:

• Minimal LLM Group
• Minimal VisionLM Group
• Minimal DALL-E Image Gen Group
• Minimal Flux Image Gen Group
• Minimal SD3 Image Gen Group
• Minimal T2I API Call Group

Example Workflow Notes

The sample workflow is meant to show minimal wiring for each API category rather than provide a polished end-to-end app.

Some nodes referenced in minimal_api_group.json come from other ComfyUI packs, such as text box or URL-loading helper nodes. If those nodes are missing in your setup, you can still use the BillBum API nodes themselves and replace the helper nodes with equivalents available in your environment.

ComfyUI BillBum APIset Nodes

ComfyUI BillBum APIset Nodes is a custom node pack for calling external APIs from ComfyUI workflows. It focuses on OpenAI-compatible LLM and VLM endpoints, image generation APIs, provider-specific image and video APIs, and a small set of utility nodes for moving data between text, URLs, images, and base64 payloads.

This repository is useful if you want to:

• run text generation from chat-style LLM APIs inside ComfyUI
• caption or analyze images with vision-capable models
• call image generation APIs such as DALL-E, SD3, Flux, GPT-Image-1, Ideogram, and Recraft
• test custom image-generation endpoints without writing glue code
• convert between image tensors, URLs, and base64 data in a workflow

Workflow Patterns by Node Group

These diagrams use a GitHub-safe Mermaid style: plain flowcharts, simple node shapes, and basic decision branches. They only show the most common external connections between nodes. Internal widget settings such as model, API URL, API key, seed, size, and similar parameters are intentionally omitted.

Troubleshooting

• If the nodes do not appear in ComfyUI, make sure the repository is inside custom_nodes and restart ComfyUI after installing dependencies.
• If a node returns authentication errors, check the API key, base URL, and whether the endpoint matches the provider.
• If a request fails with model errors, confirm that the selected model is actually available on the provider you configured.
• If image or video helper nodes fail, verify that the returned URL or base64 payload format matches what the downstream node expects.
• If the sample workflow cannot load fully, install the missing third-party helper nodes or replace them with local equivalents.