README.md

---
title: 🧩 DiffuseCraft Mod (SDXL/SD1.5 Models Text-to-Image)
emoji: 🧩🖼️📦
colorFrom: red
colorTo: pink
sdk: gradio
sdk_version: 5.45.0
app_file: app.py
pinned: true
header: mini
license: mit
duplicated_from: r3gm/DiffuseCraft
short_description: Stunning images using stable diffusion.
preload_from_hub:
  - madebyollin/sdxl-vae-fp16-fix config.json,diffusion_pytorch_model.safetensors
hf_oauth: true
---

## Using this Space programmatically

You can call this Space from Python (via `gradio_client`) or from plain `curl`.

> ⚠️ Note: This README may lag behind the actual API definition shown in the Space’s “View API” page.
> If something does not work, always double-check the latest argument list and endpoint names there.

Assumptions:

- Space ID: `John6666/DiffuseCraftMod`
- You have a valid Hugging Face access token: `hf_xxx...` (read access is enough)
- Replace `hf_xxx...` with your own token

---

### 1. Python examples (`gradio_client`)

Install:

```bash
pip install gradio_client
````

#### 1.1 Synchronous API – `generate_image`

```python
from gradio_client import Client

client = Client("John6666/DiffuseCraftMod", hf_token="hf_xxx...")

status, images, info = client.predict(
    # Core text controls
    prompt="Hello!!",
    negative_prompt=(
        "lowres, bad anatomy, bad hands, missing fingers, extra digit, "
        "fewer digits, worst quality, low quality"
    ),

    # Basic generation controls
    num_images=1,
    num_inference_steps=28,
    guidance_scale=7.0,
    clip_skip=0,
    seed=-1,

    # Canvas / model / task (optional, server has defaults)
    height=1024,
    width=1024,
    model_name="votepurchase/animagine-xl-3.1",
    vae_model="None",
    task="txt2img",

    # All other arguments are optional; defaults match the UI
    api_name="/generate_image",
)

print(status)   # e.g. "COMPLETE"
print(images)   # list of image paths / URLs
print(info)     # generation metadata (seed, model, etc.)
```

#### 1.2 Streaming API – `generate_image_stream`

```python
from gradio_client import Client

client = Client("John6666/DiffuseCraftMod", hf_token="hf_xxx...")

job = client.submit(
    prompt="Hello!!",
    negative_prompt=(
        "lowres, bad anatomy, bad hands, missing fingers, extra digit, "
        "fewer digits, worst quality, low quality"
    ),
    num_images=1,
    num_inference_steps=28,
    guidance_scale=7.0,
    clip_skip=0,
    seed=-1,
    height=1024,
    width=1024,
    model_name="votepurchase/animagine-xl-3.1",
    vae_model="None",
    task="txt2img",
    api_name="/generate_image_stream",
)

for status, images, info in job:
    # You will see progress messages, intermediate previews, and the final result.
    print(status, images, info)
```

You can stop iterating once you see a `"COMPLETE"` status if you only care about the final output.

---

### 2. `curl` examples

When calling from `curl`, include your HF token; anonymous calls may be rate-limited or rejected.

```bash
export HF_TOKEN="hf_xxx..."  # your Hugging Face access token
```

The `data` field is a positional array. The order must match the function signature.
For simplicity, the examples below only send the first few arguments and rely on server defaults for the rest.

#### 2.1 Synchronous API – `generate_image`

```bash
curl -X POST "https://john6666-diffusecraftmod.hf.space/call/generate_image" \
  -H "Authorization: Bearer $HF_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "data": [
      "Hello!!",  // prompt
      "lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, worst quality, low quality", // negative_prompt
      1,          // num_images
      28,         // num_inference_steps
      7.0,        // guidance_scale
      0,          // clip_skip
      -1          // seed
      // All subsequent parameters will use their default values
    ]
  }'
```

#### 2.2 Streaming API – `generate_image_stream`

```bash
curl -X POST "https://john6666-diffusecraftmod.hf.space/call/generate_image_stream" \
  -H "Authorization: Bearer $HF_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "data": [
      "Hello!!",
      "lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, worst quality, low quality",
      1,
      28,
      7.0,
      0,
      -1
    ]
  }'
```

For full parameter coverage (all advanced options such as LoRAs, ControlNet, IP-Adapter, etc.),
refer to the Space’s “View API” page and adapt the examples above accordingly.