AI Agent Hub
Back to skills
Agnes Text-to-Image icon

Agnes Text-to-Image

Design & Media Updated 2026.08.30

Paste the following prompt into your AI chat to install this skill:

Please follow the guide at https://skillhub.cn/install/skillhub.md to install @user_c28beb7a/agnes-image-gen into your AI assistant.

About this skill

Problem

When calling image-generation APIs from local scripts or engineering workflows, common issues include unstable endpoints, unpredictable output sizes, incorrect image-to-image parameters, and temporary image URLs that expire quickly.

How It Works

This skill wraps the Agnes AI agnes-image-2.1-flash model and uses a direct HTTP API for two workflows:
- Text-to-image: generate an image from a prompt, with 1K, 2K, 3K, or 4K tiers and ratios such as 1:1 or 16:9.
- Image-to-image: pass a public URL or Data URI into extra_body.image for style transfer, background replacement, or relighting while preserving the original layout.

Key behaviors include:
- Dual-endpoint fallback: perform a lightweight reachability check, then switch to the alternate endpoint when the primary one is unavailable, reducing network variability.
- Output control: return either url or b64_json, which is useful for quick preview, local saving, or offline processing.
- Error handling: cover common failures such as model_not_found, rate_limit_exceeded, and invalid_image_format, with client timeouts suggested at 60 seconds or higher.

Limits

It suits creative design, marketing assets, and product visualization tasks. Save generated images promptly because URLs are temporary; private images should be converted to Base64; and for exact display resolutions, request the nearest tier and crop downstream.

Use Cases

  • Marketing operations need a tech-style launch poster with blue tones for an AI assistant release.
  • A designer converts a product photo into cyberpunk style while keeping the subject layout for social testing.
  • A front-end engineer generates a 16:9 banner at 2K, then downloads and crops it for a demo page.
  • A content editor creates a holiday card with clear “Happy New Year” text using optimized Chinese prompts.

Best For

  • Visual designers who need to turn prompts into deliverable launch posters or social assets.
  • E-commerce operations who want to restyle, rebackground, or relight product photos while keeping the item visible.
  • Application engineers who integrate image-generation APIs and handle endpoints, sizes, and Base64 responses.
  • Front-end engineers who generate banners or cards and save them into project asset folders.