Qwen Image 2 ProSAFE

inference.sh Agent skills for using our API to give your agents access to hundreds of apps and other agents

The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.

---
name: qwen-image-2-pro
description: "Generate images with Alibaba Qwen-Image-2.0-Pro via inference.sh CLI. Professional text rendering, fine-grained realism, enhanced semantic adherence. Ideal for posters, banners, and text-heavy designs. Triggers: qwen image pro, qwen-image-pro, qwen 2 pro, alibaba image pro, dashscope pro, professional text rendering"
allowed-tools: Bash(belt *)
---

> **Install the belt CLI skill:** `npx skills add belt-sh/cli`

# Qwen-Image Pro - Professional Image Generation

Generate images with Alibaba Qwen-Image-2.0-Pro via [inference.sh](https://inference.sh) CLI. Best for professional text rendering and complex designs.

![Qwen-Image-2.0-Pro](https://cloud.inference.sh/app/files/u/4mg21r6ta37mpaz6ktzwtt8krr/01kjtkdzbxw3ma6wx4j1e42yry.jpeg)

## Quick Start

> Requires inference.sh CLI (`belt`). [Install instructions](https://raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md)

```bash
belt login

belt app run alibaba/qwen-image-2-pro --input '{"prompt": "Poster with title \"Welcome!\" in bold blue text"}'
```


## Pro Model Capabilities

- **Professional Text Rendering**: Multi-line and paragraph-level text with fine-grained detail
- **Fine-grained Realism**: Better textures and photorealistic scenes
- **Stronger Semantic Adherence**: More accurately follows complex prompts
- **Complex Designs**: Ideal for text + image combinations

## Examples

### Basic Text-to-Image

```bash
belt app run alibaba/qwen-image-2-pro --input '{
  "prompt": "A futuristic cityscape at sunset with flying cars"
}'
```

### Text-Heavy Poster

```bash
belt app run alibaba/qwen-image-2-pro --input '{
  "prompt": "Healing-style hand-drawn poster featuring three puppies playing with a ball. The main title \"Come Play Ball!\" is prominently displayed at the top in bold, blue cartoon font. Below, the subtitle \"Join the Fun!\" appears in green font.",
  "width": 1024,
  "height": 1536,
  "prompt_extend": false
}'
```

### Marketing Banner

```bash
belt app run alibaba/qwen-image-2-pro --input '{
  "prompt": "Professional marketing banner for summer sale. Large text \"SUMMER SALE\" in white on gradient sunset background. \"50% OFF\" in yellow below. Clean, modern design.",
  "width": 1920,
  "height": 1080,
  "prompt_extend": false,
  "negative_prompt": "blurry text, distorted text, low quality"
}'
```

### Multiple Variations

```bash
belt app run alibaba/qwen-image-2-pro --input '{
  "prompt": "Minimalist logo design for a coffee shop called \"Bean & Brew\"",
  "num_images": 4
}'
```

### Image Editing (Style Transfer)

```bash
belt app run alibaba/qwen-image-2-pro --input '{
  "prompt": "Make the person from Image 1 wear the outfit from Image 2",
  "reference_images": [
    {"uri": "https://example.com/person.jpg"},
    {"uri": "https://example.com/outfit.jpg"}
  ],
  "num_images": 2
}'
```

### Reproducible Generation

```bash
belt app run alibaba/qwen-image-2-pro --input '{
  "prompt": "Abstract geometric art in blue and gold",
  "seed": 12345
}'
```

## Input Options

| Parameter | Type | Description |
|-----------|------|-------------|
| `prompt` | string | **Required.** What to generate or edit (max 800 chars) |
| `reference_images` | array | Input images for editing (1-3 images) |
| `num_images` | integer | Number of images to generate (1-6) |
| `width` | integer | Output width in pixels (512-2048) |
| `height` | integer | Output height in pixels (512-2048) |
| `watermark` | boolean | Add "Qwen-Image" watermark |
| `negative_prompt` | string | Content to avoid (max 500 chars) |
| `prompt_extend` | boolean | Enable prompt rewriting (default: true) |
| `seed` | integer | Random seed for reproducibility (0-2147483647) |

**Size constraint:** Total pixels must be between 512×512 and 2048×2048.

## Output

| Field | Type | Description |
|-------|------|-------------|
| `images` | array | The generated or edited images (PNG format) |
| `output_meta` | object | Metadata with dimensions and count |

## Text Rendering Tips

For best text results with the Pro model:

1. **Use quotes** around exact text: `"Title: \"Hello World!\""`
2. **Specify font details**: color, style, size, position
3. **Disable prompt_extend**: Set `prompt_extend: false` for precise control
4. **Use negative prompts**: `"blurry text, distorted text, low quality"`

**Example prompt structure:**
```
Poster with the title "GRAND OPENING" in large red serif font at the top center.
Below, the date "March 15, 2024" in smaller black text.
Background: elegant gold and white gradient.
Style: professional, clean, modern.
```

## Recommended Negative Prompt

```json
{
  "negative_prompt": "low resolution, low quality, deformed limbs, deformed fingers, oversaturated, waxy, no facial details, overly smooth, AI-like, chaotic composition, blurry text, distorted text"
}
```

## Sample Workflow

```bash
# 1. Generate sample input to see all options
belt app sample alibaba/qwen-image-2-pro --save input.json

# 2. Edit the prompt
# 3. Run
belt app run alibaba/qwen-image-2-pro --input input.json
```

## Python SDK

```python
from inferencesh import inference

client = inference()

# Text-heavy poster
result = client.run({
    "app": "alibaba/qwen-image-2-pro",
    "input": {
        "prompt": "Poster with title \"Welcome!\" in bold blue text at top",
        "width": 1024,
        "height": 1536,
        "prompt_extend": False
    }
})
print(result["output"])

# Stream live updates
for update in client.run({
    "app": "alibaba/qwen-image-2-pro",
    "input": {
        "prompt": "Professional product photography of a watch"
    }
}, stream=True):
    if update.get("progress"):
        print(f"progress: {update['progress']}%")
    if update.get("output"):
        print(f"output: {update['output']}")
```

## Related Skills

```bash
# Standard Qwen-Image (faster, general use)
npx skills add inference-sh/skills@qwen-image

# Full platform skill (all apps)
npx skills add inference-sh/skills@infsh-cli

# All image ge

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

What the source does

Filesystem

none-observed

Network

none-observed

Shell

none-observed

Dependencies

pinned

Secrets in source

none-found

Findings (1)

Gates applied: no_behavioural_pass, no_license.