Services/Image Generation

Image Generation

AI image generation and editing — create from scratch or modify reference images with optional branding. Choose your model: Nano Banana 2 (Google) or gpt-image-2 (OpenAI).

OVERVIEW

Image Generation is a two-mode service. Pick your model in setup — Nano Banana 2 (Google) or gpt-image-2 (OpenAI) — bring your own API key, and switch between them anytime. In create mode, describe the visual you need — a hero banner, social graphic, ad creative, infographic, or product shot — and Agentcy builds an optimized structured prompt that produces the best results. In modify mode, provide one or more reference images and describe what you want changed — swap an outfit, change a background, restyle the image — and the AI applies your modifications while preserving the original subject.

Brand colors, fonts, and logo are opt-in. They are only applied when you explicitly ask for branded content (e.g., "use my brand colors", "add my logo"). Having a domain does not automatically trigger branding. The service works with or without a domain — for general images, just describe what you want.

Generated images are available in 1K, 2K, or 4K resolution across multiple aspect ratios (1:1 for social, 16:9 for banners, 9:16 for stories, and more). Images are stored temporarily for 30 days and returned as downloadable URLs.

EXAMPLE QUESTIONS

16 SAMPLES

“Create a hero banner for my website”

“Generate a social media post image for a holiday sale”

“Design a YouTube thumbnail for my latest video”

“Make a product photo for my coffee brand”

“Create an Instagram story graphic with bold typography”

“Generate a Facebook ad creative for our summer collection”

“Design a professional header image for my blog”

“Create a Pinterest pin for my recipe post”

“Generate a promotional banner for Black Friday”

“Create an infographic about our quarterly performance”

“Make a branded image using my company colors and logo”

“Create a 4K hero shot for my landing page”

“I want an image like this but with a different outfit”

“Edit this photo to change the background to a beach”

“Transform this photo into a watercolor painting style”

“Take this image and make the person wear a hockey jersey”

WHAT YOU CAN ASK ABOUT

6 EXAMPLES
Generated PNG images (1K, 2K, or 4K resolution)
Multiple aspect ratios: 1:1 (social), 16:9 (banners), 9:16 (stories), 3:4 (Pinterest), 4:3, 3:2, 2:3
Create mode: from-scratch image generation from text descriptions
Modify mode: edit reference images — change outfits, backgrounds, styles while preserving subjects
Opt-in branding: brand colors, fonts, and logo placement when explicitly requested
Reference image support (up to 5 images — photos, logos, style references)

REQUIREMENTS

Google Account
Not required
API Key
Provide your own API credentials
External Subscription
Not required
Domain Configuration
None — zero-config service

SETUP

5 STEPS
  1. 01

    Get an API key for your chosen model — Gemini (ai.google.dev) for Nano Banana 2, or OpenAI (platform.openai.com) for gpt-image-2

  2. 02

    In the Agentcy portal, go to Services

  3. 03

    Enable 'Image Generation'

  4. 04

    Pick your model and enter the matching API key (your OpenAI organization must be verified to use gpt-image-2)

  5. 05

    Start generating images via the marketing_data tool

DATA FRESHNESS

Real-time generation. Each request produces a new image on demand. No cached data.

TIPS

4

Brand colors and logo are only applied when you explicitly ask — having a domain does not auto-trigger branding

In modify mode, describe what to CHANGE, not what the image already shows — the AI can see the reference

Image generation costs are on your own API usage (Gemini or OpenAI, depending on your chosen model) — not included in the Agentcy subscription

AI text rendering in images can be imperfect — keep text overlays simple (1-2 lines)

WORKS WELL WITH

3 SERVICES

CONNECT VIA

POPULAR CLIENTS

View all integrations →

v1.0Need help?