Generate images from text prompts using fal.ai. Use when the user asks to create, generate, or make an image from a text description. Supports multiple aspect ratios, resolutions up to 4K, and sync mode for reliable results.
Resources
2Install
npx skillscat add bluewaves-creations/bluewaves-skills/image-generator Install via the SkillsCat registry.
This skill generates images from text prompts using fal.ai's Gemini 3 Pro model, supporting various aspect ratios, resolutions up to 4K, and sync mode for reliable results. It solves the problem of creating visual content from textual descriptions without requiring local model deployment or complex setup. Developers should use it when an application needs to convert user-provided text descriptions into images, particularly when high resolution, multiple output formats, or web-grounded generation are required.
Image Generator
Generate high-quality images from text prompts using Google's Gemini 3 Pro model via fal.ai.
See references/fal-api.md for setup, Python patterns, and error handling.
API Endpoint
fal-ai/gemini-3-pro-image-preview
Parameters
Required
prompt(string): Text description of the image to generate
Optional
| Parameter | Type | Default | Options |
|---|---|---|---|
num_images |
integer | 1 | 1-4 |
aspect_ratio |
string | "1:1" | "21:9", "16:9", "3:2", "4:3", "5:4", "1:1", "4:5", "3:4", "2:3", "9:16" |
output_format |
string | "png" | "jpeg", "png", "webp" |
resolution |
string | "1K" | "1K", "2K", "4K" |
sync_mode |
boolean | false | Returns data URI directly (recommended for reliability) |
enable_web_search |
boolean | true | Grounds generation in current web data for styles and references |
safety_tolerance |
string | "6" | Always set to "6" for maximum creative freedom |
seed |
integer | - | Fix seed for reproducible results |
limit_generations |
boolean | false | Restricts to 1 image per prompt round |
CLI Script
python3 scripts/fal_generate.py \
--endpoint image \
--prompt "A mountain landscape at golden hour" \
--aspect-ratio 16:9 \
--resolution 2K \
--output landscape.pngTips
- Set
enable_web_search: Trueto ground generation in real-world styles, brands, and visual references - Always include
safety_tolerance: "6"for maximum creative flexibility - Use
sync_mode: Truefor the most reliable workflow (no separate download) - Be specific in prompts: include lighting, mood, composition, and style descriptors
- Match aspect ratio to use case: 16:9 for landscapes, 9:16 for portraits, 1:1 for social media
- Higher resolution (4K) produces more detail but takes longer to generate
- Use
seedto reproduce a specific result