gemini-image_skill

Overview

This skill generates high-quality images from text prompts using the Gemini 3 Pro model via fal.ai. It supports multiple aspect ratios and resolutions up to 4K, returning URLs or data URIs for immediate use. Use it when you need photorealistic, stylized, or concept imagery from brief descriptions.

How this skill works

Send a POST request with a prompt and optional parameters like num_images, aspect_ratio, resolution, and output_format to the fal.ai Gemini 3 Pro endpoint. The service returns one or more image objects with file names, content types, and accessible URLs (or embedded data URIs when sync_mode is true). You can adjust generation behavior with enable_web_search and limit_generations flags.

When to use it

• Create concept art, mood boards, or visual drafts from text descriptions
• Generate social media images with specific aspect ratios (16:9, 9:16, 1:1, etc.)
• Produce high-resolution assets (1K, 2K, 4K) for presentations or mockups
• Quickly iterate multiple variations by requesting multiple images per prompt
• Embed generated images directly in apps using data URIs for synchronous workflows

Best practices

• Be specific in prompts: include subject, style, lighting, mood, and color palette
• Choose aspect_ratio to match the target display (16:9 for landscapes, 9:16 for vertical)
• Reserve 4K only when needed; higher resolutions increase generation time and cost
• Handle errors and rate limits: retry on 500/429 with exponential backoff
• Keep FAL_KEY secure and validate it to avoid 401 Unauthorized responses

Example use cases

• Generate a 2K landscape banner for a website with aspect_ratio 16:9
• Create a portrait-oriented character illustration using aspect_ratio 9:16
• Produce four concept variations at once by setting num_images to 4
• Request a 4K, highly detailed product mockup for marketing material
• Embed a synchronous data URI image in an email or chat message for instant preview