How To Generate 1920×1080 AI Images In Bulk

by (AI-Assisted)
Edited, reviewed and fact-checked by
Published

To generate 1920×1080 AI images in bulk, prepare a CSV of text prompts, set the image Width to 1920 pixels and Height to 1080 pixels, and run the prompt list through Bulk AI Image Creator: Batch Generative Fill in Photoshop. The plugin processes the prompts sequentially and automatically exports either the first generated variation or every variation Photoshop returns.

This gives you one batch-wide size setting instead of relying on each prompt to request the right dimensions. It is particularly useful when you need a collection of 16:9 images for video thumbnails, widescreen backgrounds, or landscape artwork concepts.

Here is how to configure the batch, organize the outputs, and use this specific size effectively.

1. Start With Pixel Dimensions, Not Just a “16:9” Prompt

1920×1080 means an image that is 1,920 pixels wide and 1,080 pixels tall. Its aspect ratio is 16:9, making it a landscape canvas.

The distinction between dimensions and aspect ratio matters. An image can have a 16:9 shape without containing exactly 1920×1080 pixels. Asking an AI model for a “widescreen image” describes the composition you want; it does not replace configuring the document’s pixel dimensions.

For a consistently sized batch, use the plugin’s Width and Height controls to establish the canvas. Use your prompts to describe what belongs inside that canvas: the subject, setting, lighting, visual style, and placement of important elements.

For example, a thumbnail background might need a subject on the right and clear space on the left. A landscape art image might need a balanced composition across the full width. Both can use the same 1920×1080 settings while requiring very different prompts.

2. Open Bulk AI Image Creator Inside Photoshop

Bulk AI Image Creator: Batch Generative Fill is a third-party Photoshop plugin available through Adobe Exchange or Adobe’s Creative Cloud plugin marketplace.

  1. Find the plugin in Adobe’s plugin marketplace and acquire it.
  2. Allow Creative Cloud to install it.
  3. Open or restart Photoshop if required.
  4. Open the installed plugin through Photoshop’s Plugins menu.

The plugin automates Photoshop’s own Generative Fill functionality. It does not generate images through an independent external API account.

Once it is open, the main controls let you select your prompt CSV, enter the image dimensions, choose an AI model, configure variation exports, and select an output folder. You will also need to acknowledge Generative Credits usage before generation begins.

3. Prepare a CSV With One Image Prompt Per Row

Your CSV is the list of images you want Photoshop to generate. Its structure is deliberately simple: put TEXT PROMPTS in cell A1, then place your prompts in Column A beginning at A2.

For a small landscape-image batch, the cells could contain:

That example creates four prompt-generation operations. You do not need separate columns for dimensions, model names, or export settings. Those are configured once in the plugin and apply to the entire batch.

Keep prompt data in Column A only. Unexpected non-empty data in additional columns causes validation to fail. Blank rows beneath the header are ignored, and each usable prompt must contain at least one letter or number.

If a prompt contains commas, keep it inside a quoted CSV field so the commas do not split it into additional columns. For example, this is one prompt value:

"A cinematic mountain cabin at sunrise, warm window light, mist in the valley, wide landscape composition"

Select the finished file using CSV File of Text Prompts To Use.

4. Set Width to 1920 and Height to 1080

In the Image Size controls, enter:

Width and Height establish the pixel dimensions. DPI is a separate document-resolution setting. Leaving DPI at its default of 300 does not turn a 1920×1080 image into a larger pixel image or add detail.

For screen-based work such as video thumbnails, the pixel dimensions are the important size specification. For printing, the relationship between pixels and physical inches also matters, which we will address in the wall-art example below.

The same Width, Height, and DPI apply to every prompt in the operation. If part of your prompt library needs a portrait or square canvas, put those prompts in a separate batch rather than expecting individual CSV rows to change the dimensions.

1920×1080 is within the plugin’s accepted numerical range. Generation still depends on Photoshop and the selected model, so a short test batch is worth running before committing a large prompt list.

5. Choose One AI Model and Any Shared Reference Images

Choose a model from the AI model dropdown. The default is Gemini 3 (Nano Banana Pro), and the plugin currently provides eight model choices.

One selected model is used for the entire batch. If you want to compare how two models interpret the same 1920×1080 prompt list, run the CSV in separate operations with a different model selected each time. Including the model name in your output filenames will help keep the results identifiable.

Reference images are optional. They can be useful when you want a batch to receive common visual guidance, such as an existing illustration style or a recurring visual subject.

The plugin’s reference-image workflow supports Gemini 3 (Nano Banana Pro), Gemini 3.1 (Nano Banana 2), Gemini 2.5 (Nano Banana), FLUX2 Pro, and Flux Kontext Pro.

To use references, enable Attach Reference Images and select a folder containing one or two supported reference-image files, such as JPEG or PNG files. The same references are supplied with every prompt in the batch. More than two supported reference images in the folder will fail validation.

You do not need to classify the files into separate reference categories. Their interpretation is handled by Photoshop and the selected model; treat them as guidance rather than a guarantee that every output will look identical.

6. Decide Whether Each Prompt Should Export One Image or Every Variation

A prompt-generation operation and an exported image file are not necessarily the same thing. Photoshop can return multiple variations for a single prompt.

Export All Generated Image Variations is enabled by default. With it enabled, the plugin determines how many variations Photoshop actually produced and exports each one as a separate file.

If you want one exported image per prompt, disable this setting. The plugin then exports only the first generated variation. It does not choose the best variation for you or prevent Photoshop from internally producing alternatives.

For example, a CSV containing 50 valid prompts should produce 50 exported files in first-variation-only mode, assuming every generation succeeds. With all variations enabled, the same CSV may produce more than 50 files. The actual total depends on the variations returned by Photoshop and the selected model.

For thumbnail concepts, exporting every variation gives you more compositions to review. For a tightly controlled asset list, exporting only the first variation keeps the output count closer to the number of prompts.

7. Configure Exports and Filenames Before You Generate

Select Export Folder To Use and choose the format appropriate to your next step. JPEG is the default and a practical starting point for many screen-image workflows. PNG is useful when you want a lossless image output.

JPEG and PNG expose a user-facing Quality setting from 0 through 100. For JPEG, this controls the saving-quality setting. For PNG, it controls compression rather than making the image lossy: a higher displayed Quality means lower compression.

Use the filename builder to make the batch identifiable without opening every file. The default structure includes the prompt and model. For this workflow, adding dimensions is especially useful:

{text-prompt}-{AI-model-used}-{width}x{height}

A filename could look like:

mountain-cabin-at-sunrise-gemini-3-nano-banana-pro-1920x1080.jpg

The width and height filename components use the actual Photoshop document dimensions. This is useful when you later have several batches at different sizes.

Prompt-based filename text is sanitized and may be shortened, so it will not necessarily preserve your complete original prompt. Multiple variations receive duplicate-safe numbered names instead of overwriting one another.

8. Apply the Workflow to YouTube Thumbnails and Landscape Wall Art

YouTube Thumbnail Backgrounds for a Series of Videos

1920×1080 has the 16:9 aspect ratio used for YouTube thumbnails. A practical use is generating a collection of thumbnail backgrounds or concept images for an upcoming video series.

Start by assigning one prompt to each planned video. Rather than requesting a vague “YouTube thumbnail,” describe the visual scene and where the subject should appear.

For a camping video, a prompt might be:

"A cinematic campsite beside a mountain lake at sunrise, a bright orange tent on the right side, uncluttered pale sky on the left for a title overlay, strong subject separation, wide landscape composition"

For a technology video, another row could contain:

"A dramatic close-up of a futuristic desktop computer on the right side of a dark studio scene, blue accent lighting, simple negative space on the left for a title overlay, wide landscape composition"

Set the batch to 1920×1080, select your model, and enable all-variation export if you want several compositions to evaluate for each video. If you have suitable visual references, use a compatible model and attach the same one or two references across the batch.

After generation, review which images leave usable space for titles and which subjects remain recognizable at a small viewing size. Add final title text and other thumbnail elements in Photoshop as a separate design step. The plugin generates and exports the imagery; it does not automatically assemble a finished thumbnail layout.

Keep prompt wording consistent where it matters, such as subject placement or lighting, while changing the scene for each video. Check YouTube’s current upload requirements before publishing your finished thumbnails.

Landscape Wall-Art Concepts and Small Prints

A 1920×1080 batch can also help you develop a coordinated collection of landscape wall-art concepts. For example, you could create a series of coastal scenes, minimalist mountain illustrations, or abstract earth-tone compositions.

Write one prompt per artwork while repeating the collection’s visual direction:

Generate the set at 1920×1080 and export every variation if you want to select the strongest pieces for the collection. A shared reference image can provide additional visual guidance with a compatible model. PNG or TIFF may be useful output choices for subsequent artwork review and preparation.

There is an important distinction between creating artwork concepts and creating large print-ready posters. At 300 DPI, 1920×1080 pixels corresponds to 6.4×3.6 inches. Changing the DPI value does not create additional pixels.

For a 16×9-inch print at 300 DPI, the corresponding pixel dimensions would be 4800×2700. If your print-on-demand provider requires a larger file, use the 1920×1080 batch for concept selection, then prepare an appropriately sized production file. A separate generation batch can be configured with larger target dimensions, subject to Photoshop and model constraints.

Also confirm the final product’s aspect ratio. A 16:9 image fits a 16:9 landscape format, but not every wall-art or poster size shares that ratio. Other formats may require cropping or a differently sized canvas, along with any bleed or margin requirements specified by the print provider.

9. Run a Small Test, Then Generate the Full Prompt List

Before starting a large batch, create a short test CSV containing two or three representative prompts. Use the same 1920×1080 dimensions, model, references, export format, and filename structure you intend to use for production.

  1. Select the test CSV and confirm your image-size settings.
  2. Check the model, reference-image folder if enabled, and variation-export setting.
  3. Select the output folder, format, Quality, and filename components.
  4. Click Generate My Images.
  5. Accept the Generative Credits acknowledgement and confirm the operation.
  6. Review the exported images before selecting the full production CSV.

The plugin validates the operation before generation begins. Its confirmation step requires explicit acknowledgement of Generative Credits usage. Actual consumption depends on Adobe’s policies, the selected model, and the generation operation; the plugin does not provide an exact credit-total prediction.

During processing, each valid prompt is submitted through Photoshop. The plugin waits for the generation, exports the selected variation or variations, clears the completed prompt’s variations, and advances to the next prompt.

The progress count tracks prompt-generation operations, not the total number of exported variations. A batch with 50 valid prompts is therefore tracked as 50 operations even if it exports more files.

You can request cancellation through Cancel Operation or the Escape key while the progress interface is active. Cancellation is handled most reliably between Generative Fill operations rather than instantly interrupting an active Photoshop request.

For the test outputs, check pixel dimensions, composition, export appearance, and filenames. If the setup is right, select the full CSV and run it with the same settings. Afterward, review the images for their intended use; consistent dimensions do not eliminate the need for creative selection.

10. Save the 1920×1080 Setup for Future Projects

If you regularly create widescreen imagery, save the working configuration as a Saved Operation. Give it a descriptive name such as “1920×1080 Thumbnail Backgrounds” or “1920×1080 Landscape Art Concepts.”

A Saved Operation can retain the dimensions, DPI, model, reference-image settings, variation-export setting, export format, Quality, filename components, and references to the selected CSV and folders.

For the next project, load the operation, select a different CSV or output folder where needed, and check the restored settings. This makes 1920×1080 a repeatable production configuration rather than a size you have to remember and enter again every time.

Saved Operations retain access references to files and folders, not an embedded copy of your prompt CSV. If a resource has moved or become unavailable, you may need to select it again.

The essential workflow is straightforward: prepare one prompt per CSV row, set Width to 1920 and Height to 1080, choose your generation and export settings, and let Bulk AI Image Creator automate the prompt-by-prompt work inside Photoshop. Start with a small test, then use the same configuration to produce a consistently sized collection for your next thumbnail series, widescreen asset library, or landscape-art project.