Generate an Image via BlockEden OpenAI Proxy

You are helping the user generate an AI image using the BlockEden OpenAI proxy, which is fully compatible with the OpenAI Images API (DALL-E 3 / DALL-E 2 / gpt-image-1).

Prerequisites

The BLOCKEDEN_ACCESS_KEY environment variable must be set. If it's missing, tell the user:

Please set the BLOCKEDEN_ACCESS_KEY environment variable before using this skill.

How to generate an image

Use the bundled script at scripts/generate_image.py (relative to this SKILL.md):

python <skill_dir>/scripts/generate_image.py "<prompt>" [options]

Options

| Flag | Default | Choices | Notes | |------|---------|---------|-------| | --reference-image | None | any URI | Reference image URL to base generation on. Uses gpt-image-1 via /v1/responses. | | --size | 1024x1024 | 1024x1024, 1792x1024, 1024x1792 | Landscape: 1792x1024, Portrait: 1024x1792 | | --quality | standard | standard, hd | hd gives finer detail, costs more (text-only mode only) | | --model | dall-e-3 | dall-e-3, dall-e-2 | Model for text-only generation. Ignored when --reference-image is set. | | --style | vivid | vivid, natural | DALL-E 3 only. vivid = dramatic, natural = realistic | | --output-dir | . | any path | Directory where the PNG will be saved |

Example: text-only generation

python /path/to/skill/scripts/generate_image.py \
  "a cozy coffee shop on a rainy day, impressionist painting style" \
  --size 1792x1024 \
  --quality hd \
  --style natural \
  --output-dir ~/Desktop

Example: reference image generation

python /path/to/skill/scripts/generate_image.py \
  "same composition but set at night with neon lights" \
  --reference-image "https://example.com/my-photo.jpg" \
  --output-dir ~/Desktop

After generation

Text-only mode output

{
  "filepath": "/path/to/image_1234567890.png",
  "revised_prompt": "...",
  "model": "dall-e-3",
  "size": "1024x1024",
  "quality": "standard",
  "style": "vivid"
}

Reference image mode output

{
  "filepath": "/path/to/image_1234567890.png",
  "model": "gpt-image-1",
  "reference_image": "https://example.com/my-photo.jpg",
  "prompt": "same composition but set at night with neon lights"
}

Once the image is saved:

1. Tell the user the file path so they can open it.
2. Show the revised_prompt if DALL-E rewrote it — this helps the user refine future prompts.
3. Offer to regenerate with different parameters if the user wants to iterate.

Interpreting the user's request

• Extract the core visual description from what the user says and use it as the prompt.
• If the user provides a reference image URL/URI, pass it via --reference-image. The model switches automatically to gpt-image-1.
• If the user specifies an aspect ratio (wide, tall, square), map it to the appropriate --size.
• If the user says "high quality", "detailed", or "HD", add --quality hd (text-only mode only).
• If the user says "realistic" or "photorealistic", use --style natural.
• If the user says "artistic", "dramatic", or "vivid", use --style vivid (the default).
• If the user provides an output location, pass it via --output-dir.
• If no output location is given, save to the current working directory.

Prompt crafting tips

Good image prompts are specific and visual. If the user's request is vague, enhance it slightly:

• Add a medium/style if none given (e.g., "digital art", "watercolor", "photorealistic")
• Add lighting or atmosphere if it feels generic (e.g., "golden hour lighting", "moody atmosphere")
• Don't over-engineer — keep the user's intent central