The short answer
The prompts below are free to copy. In the Gemini app, Google says images can be downloaded at 1K resolution without an AI plan; a Google AI plan allows 2K downloads. No-plan access is still subject to changing usage limits, not an unlimited image allowance. Availability depends on supported countries and languages. For personal accounts, Google’s image help page says users must be 13 or older to generate images and 18 or older to generate and edit images. Work or school accounts may have different restrictions. (Google’s image-generation help; Gemini app limits)
That app access does not establish that production API image generation is free. Google documents a stable image-generation model, Nano Banana 2.1 (gemini-nano-banana-2.1), and describes free and paid API tiers. But the official pricing information retrieved for this guide does not establish this model’s image-output rate, billing unit, or eligibility for the free tier. So an exact API cost per image—or a dependable workload estimate—remains unconfirmed here. (Nano Banana 2.1 model documentation; Gemini API pricing)
| Route | What’s confirmed | Budget caveat |
|---|---|---|
| Gemini app, no AI plan | Google says 1K image downloads are available without an AI plan. | Standard usage limits apply and may change; Google does not give a fixed image count in the cited limits guidance. |
| Gemini app, Google AI plan | Google says plan holders can download images at 2K. | A plan offers expanded access, not a promise of unlimited generation. |
| Production API | Nano Banana 2.1 (gemini-nano-banana-2.1) is documented as a stable image-generation model. Google’s API pricing page describes free access for certain models and paid access for production applications. |
The retrieved official pricing text does not confirm this image model’s free-tier eligibility or its model-specific image rate and billing unit. Verify those terms before estimating spend. |
Free-to-copy Gemini image prompts
Copying these prompts costs nothing; generating an image is a separate product-access and usage-limit question. The examples below are original editorial prompts informed by Google’s guidance: specify the subject, composition, action, location, and style; when editing an existing image, give direct, specific instructions. (Google’s prompt guidance)
1. Generate a new image
Create a warm editorial photograph of a small independent bookstore on a rainy evening. Frame it from across the street, with amber light glowing through the front windows and reflections on the sidewalk. Include a bicycle near the entrance. Use natural colors and realistic photography; do not include readable signs or lettering.
Customize the subject, framing, setting, lighting, and style. Asking for no readable lettering is an instruction, not a guarantee of perfect text rendering.
2. Edit an uploaded image
In the uploaded photo, change only the background to a quiet beach at sunset. Keep the person’s clothing, pose, and expression unchanged. Match the light on the person to the warm sunset and keep the result realistic.
This is an editing prompt: upload an image first, then specify what to change and what to preserve. Preservation instructions are requests, not guarantees. Google’s current help page lists a higher age requirement for image editing than for image generation; check the eligibility details above before trying an edit. (Google’s image-generation help)
3. Continue a character in the same conversation
Show the same small, blue-feathered bird from the previous image perched on a wooden windowsill during a gentle snowfall. Keep its round shape, yellow scarf, and curious expression. Use a soft storybook illustration style.
Google’s published guidance uses follow-up prompts in the same conversation to try to retain character details, while noting that character consistency can still be imperfect. (Google’s prompt guidance)
What to check before budgeting for API use
Google’s official pricing page distinguishes a free tier—with access to certain models—from paid access intended for production applications. It also says that content submitted on the free tier may be used to improve Google products, while paid-tier content is not. These are tier-level terms; they do not, by themselves, confirm Nano Banana 2.1’s free-tier eligibility or price. (Gemini API pricing)
For a production budget, confirm the current model-specific price and billing unit in Google’s pricing documentation and calculate the image-output portion from those rates. Also account for any separately billed inputs or other request features if Google lists them. The official model page identifies gemini-nano-banana-2.1 as stable and supporting image generation, but the pricing information retrieved here does not supply the rate needed to calculate a reliable per-image or workload total. Do not use the Gemini app’s no-plan access as evidence that API calls are free. (Nano Banana 2.1 model documentation; Gemini API pricing)
Prompt limitations
Google’s August 2025 prompting guidance notes that stylization may be inconsistent, generated text may be misspelled, character consistency may fail, and requested aspect ratios may not be maintained. Treat prompts as instructions to try—not as a promise of a particular output. (Google’s image-generation prompting guidance)
Sources
Generate & edit images with Gemini Apps
Gemini Apps limits & feature availability