### Install ComfyKit Custom Nodes

Source: https://context7.com/bobmagicii/comfykit-custom-nodes/llms.txt

Provides the shell commands to clone the repository into the ComfyUI custom_nodes directory.

```bash
cd ComfyUI/custom_nodes/
git clone https://github.com/bobmagicii/comfykit-custom-nodes.git
```

--------------------------------

### LoraStackFiveSimple: Blend up to 5 LoRAs with synchronized weights

Source: https://context7.com/bobmagicii/comfykit-custom-nodes/llms.txt

The LoraStackFiveSimple node blends up to 5 LoRA adapters into a base model and CLIP encoder, synchronizing their weights. It automatically extracts trigger words from associated .txt files and returns the blended model, CLIP, and a comma-separated string of all trigger words. This node simplifies the process of applying multiple LoRAs in ComfyUI workflows.

```python
# ComfyUI Node Configuration
# Category: bobmagicii
# Display Name: "LoRA S5"

# Node Input Specification
INPUT_TYPES = {
    "required": {
        "model": ("MODEL",),           # Base Stable Diffusion model
        "clip": ("CLIP",),             # CLIP text encoder
        "lora1_mdl": (lora_list,),     # First LoRA file selection (or "None")
        "lora1_str": ("FLOAT", {"default": 1.0, "min": -16.0, "max": 16.0, "step": 0.1}),
        "lora2_mdl": (lora_list,),     # Second LoRA
        "lora2_str": ("FLOAT", ...),
        "lora3_mdl": (lora_list,),     # Third LoRA
        "lora3_str": ("FLOAT", ...),
        "lora4_mdl": (lora_list,),     # Fourth LoRA
        "lora4_str": ("FLOAT", ...),
        "lora5_mdl": (lora_list,),     # Fifth LoRA
        "lora5_str": ("FLOAT", ...)
    }
}

# Node Output Specification
RETURN_TYPES = ("MODEL", "CLIP", "STRING")
RETURN_NAMES = ("MODEL", "CLIP", "TRIGGERS")

# Example workflow usage:
# 1. Connect a checkpoint loader's MODEL and CLIP outputs to this node
# 2. Select up to 5 LoRAs from the dropdown (set unused slots to "None")
# 3. Adjust strength values (1.0 = full effect, 0.0 = no effect)
# 4. Connect MODEL and CLIP outputs to KSampler
# 5. Connect TRIGGERS output to prompt text for automatic trigger word insertion

# Trigger word files should be placed alongside LoRAs:
# models/loras/my_style.safetensors
# models/loras/my_style.txt  <- Contains trigger words (first line only)

```

--------------------------------

### TypecasterModel: Explicitly recast MODEL type for workflow stability

Source: https://context7.com/bobmagicii/comfykit-custom-nodes/llms.txt

The TypecasterModel node passes through a MODEL input while explicitly declaring its output type as MODEL. This is crucial for resolving ComfyUI workflow initialization issues, particularly when connecting to nodes that accept dynamic 'anything' types, ensuring stable connections on first load.

```python
# ComfyUI Node Configuration
# Category: bobmagicii
# Display Name: "Recast Model"

INPUT_TYPES = {
    "required": {
        "model": ("MODEL",)
    }
}

RETURN_TYPES = ("MODEL",)
RETURN_NAMES = ("MODEL",)

# Usage: Insert between any dynamic output node and a MODEL-accepting input
# Example workflow:
# [Switch Node (ANY)] --> [Recast Model] --> [KSampler (MODEL)]
#
# Without Recast: Connection may fail on first workflow load
# With Recast: Connection is explicitly typed and works immediately

```

--------------------------------

### TypecasterImage: Explicitly recast IMAGE type for image processing chains

Source: https://context7.com/bobmagicii/comfykit-custom-nodes/llms.txt

The TypecasterImage node provides explicit type declaration for IMAGE tensors, ensuring reliable connections within image processing pipelines in ComfyUI. By passing through an IMAGE input and specifying the output as IMAGE, it resolves type issues that may arise from dynamic node outputs in complex image generation or manipulation workflows.

```python
# ComfyUI Node Configuration
# Category: bobmagicii
# Display Name: "Recast Image"

INPUT_TYPES = {
    "required": {
        "image": ("IMAGE",)
    }
}

RETURN_TYPES = ("IMAGE",)
RETURN_NAMES = ("IMAGE",)

# Usage: Resolve type issues in image processing chains
# [Image Switch (ANY)] --> [Recast Image] --> [VAE Encode]
# [Image Router] --> [Recast Image] --> [Save Image]

```

--------------------------------

### TypecasterClip: Explicitly recast CLIP type for type inference issues

Source: https://context7.com/bobmagicii/comfykit-custom-nodes/llms.txt

The TypecasterClip node ensures stable connections in ComfyUI by passing through a CLIP input and explicitly declaring the output type as CLIP. It addresses type inference problems that can occur in complex workflows, particularly when dealing with dynamically typed nodes connected to CLIP encoder inputs.

```python
# ComfyUI Node Configuration
# Category: bobmagicii
# Display Name: "Recast CLIP"

INPUT_TYPES = {
    "required": {
        "clip": ("CLIP",)
    }
}

RETURN_TYPES = ("CLIP",)
RETURN_NAMES = ("CLIP",)

# Usage: Insert between dynamic nodes and CLIP-accepting inputs
# [Dynamic Loader] --> [Recast CLIP] --> [CLIP Text Encode]

```

--------------------------------

### Define Recast Conditioning Node

Source: https://context7.com/bobmagicii/comfykit-custom-nodes/llms.txt

Configures a ComfyUI node to pass through Conditioning data. It helps resolve type mismatches when using dynamic conditioning combiners.

```python
INPUT_TYPES = {
    "required": {
        "conditioning": ("CONDITIONING",)
    }
}

RETURN_TYPES = ("CONDITIONING",)
RETURN_NAMES = ("CONDITIONING",)
```

--------------------------------

### Define Recast VAE Node

Source: https://context7.com/bobmagicii/comfykit-custom-nodes/llms.txt

Configures a ComfyUI node to pass through VAE data. This is essential for maintaining type consistency in multi-VAE selection workflows.

```python
INPUT_TYPES = {
    "required": {
        "vae": ("VAE",)
    }
}

RETURN_TYPES = ("VAE",)
RETURN_NAMES = ("VAE",)
```

--------------------------------

### TypecasterLatent: Explicitly recast LATENT type for conditional routing

Source: https://context7.com/bobmagicii/comfykit-custom-nodes/llms.txt

The TypecasterLatent node ensures robust connections in ComfyUI workflows involving latent tensors by explicitly declaring the output type as LATENT. This node is essential for workflows with conditional latent routing, preventing type-related errors and ensuring smooth data flow.

```python
# ComfyUI Node Configuration
# Category: bobmagicii

```

--------------------------------

### Define Recast Latent Node

Source: https://context7.com/bobmagicii/comfykit-custom-nodes/llms.txt

Configures a ComfyUI node to pass through Latent data while explicitly declaring the output type. This is used to resolve connection issues in branching workflows where type inference fails.

```python
INPUT_TYPES = {
    "required": {
        "latent": ("LATENT",)
    }
}

RETURN_TYPES = ("LATENT",)
RETURN_NAMES = ("LATENT",)
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.