ImageGenerator is an online AI image generation and editing platform whose core purpose is to turn text descriptions or reference images into finished images, and to provide a variety of processing tools around the same image. According to the official introduction, it can be used online for free, and generated results are uniformly saved in a private image library.
[Main Capabilities]
-
Text-to-image: Describe the subject, style, composition, and lighting in natural language, and the generator turns the prompt into a complete image. The official service says no negative prompts or random seeds are needed.
-
Reference image generation: You can upload an existing image as a visual reference and change the style, scene, lighting, composition, or subject based on it, which is suitable for product variations, campaign concepts, and creative exploration.
-
Text rendering within images: Readable text can be generated in posters, packaging, and title cards, and the official service says non-Latin scripts are supported. The official service states that Nano Banana Pro and GPT Image 2 perform strongest on text effects.
-
Multi-model selection: Within the same prompt input box, GPT Image 2.5 Flare, Nano Banana Pro, GPT Image 2, and Seedream 5.0 Pro can be used. The official positioning of the models is respectively: GPT Image 2.5 Flare is suitable for rapid drafts, Nano Banana Pro is suitable for dense detail, GPT Image 2 is suitable for layouts that faithfully follow prompts, and Seedream 5.0 Pro is suitable for production-grade visuals and multi-reference image work.
-
Single-image iteration: Generate one image at a time, review the result, then adjust the prompt, or use that image as a reference for the next round of iteration.
-
Aspect ratio control: Before generation, you can choose 1:1, 16:9, 9:16, 4:3, or 3:4, and the model composes according to the ratio rather than cropping after generation.
-
Private image history: Each generation saves the prompt and the model used to a private library, where you can download full-resolution files, reuse prompts, or use the image as a reference to generate again.
-
Auxiliary editing tools: The official page mentions that existing images can be processed with background removal, upscaling, outpainting, or style changes. Each tool focuses on one workflow, and the results are likewise saved to the same private image library.
[How to Use]
No software installation is required, and no API key of your own is needed. The official service says the models run on the server side, with no download, no GPU, and no key to paste, and it can be used in any laptop or mobile browser. After registering, you receive welcome credits, and the free quota is enough to generate several images, with no payment information required. GeneBefore generation, the points consumed for this session will be displayed; after logging in, the balance will replace the display.
[Plans and Points]
The official offering provides subscription plans and one-time point packs. Subscriptions renew automatically and can be canceled at any time; one-time point packs are valid for one year. The officially listed plans include: 24,000 points (suitable for casual creators, valid for one year), 54,000 points (suitable for professionals, valid for one year), and 120,000 points (suitable for heavy users, valid for one year). These plans all include text-to-image, image-to-image, all image models, free downloads, and commercial licensing. The official offering also provides a model-based estimated points consumption table, assuming all points are used with a single model at 1K resolution; other resolutions consume different amounts.
[Use Cases]
The official description can be used for rapid sketches, densely detailed images, layouts that require faithful reproduction of prompts, production-grade visuals, and multi-reference image work; it can also be used for posters, packaging, title cards, and other scenarios that require readable text within the image, as well as product variations, campaign concepts, and creative exploration.
[Limitations and Notes]
The official FAQ mentions that the points consumed per generation depend on the model and the number of generations, with GPT Image 2 starting at 12 points. A free tier is available, but higher-fidelity models, more monthly points, and API access belong to paid plans. The specific model usage, the exact values in the points consumption table, the specific consumption at other resolutions, and the specific details of API access are not fully provided on the official page. The plan prices themselves are not yet provided in the materials.
