Create images from prompts, edit reference photos, make videos, or split a picture into transparent layers. ImageLayered brings these tools together with a shared credit balance.
Used this tool? Rate it
Used this tool? Rate it
ImageLayered is an AI creative platform for generating images, editing reference photos, exploring visual styles, removing backgrounds, extending compositions, and separating images into editable layers. It also includes video creation and editing tools. Its current scope covers images, photos, and video, so the layer-separation feature is one part of a larger creative toolbox.
Tools share a credit balance, and the cost depends on the tool and settings and is shown before generation. Individual tools have different reference limits, output formats, resolution choices, and credit costs.
The model names and quality descriptions below are the vendor's labels and claims, including the generators presented as GPT Image 2.5, Qwen Image 2.1, and Nano Banana 2.1.
ImageLayered offers several AI image generator workspaces with different reference capacities and export controls:
The image to layers tool turns a flat photo into multiple transparent RGBA layers that can be moved, recolored, or replaced. Precision Layers is positioned around named layers with preserved positions, while Quick Layers uses Qwen Image Layered for a simpler initial split. The Quick interface shows a choice of two to eight layers.
Precision mode accepts PNG or JPEG files up to 10MB. Its upload notice states “at least 512×512 total pixels” and an aspect-ratio range of 1:16–16:1. The interface lists two to thirteen outputs; the example counts the base image among the outputs.
Layer separation also includes reconstructed hidden regions. The vendor says the model reads the scene, separates elements, and rebuilds occluded areas so individual layers can be edited on their own. Rebuilt hidden regions are generated content.
The background remover has a simpler output goal. It detects subjects such as people, products, or pets and returns a transparent PNG. It accepts JPG, PNG, or WebP files up to 10MB and lists a cost of 10 credits per image.
The AI Image Extender generates surroundings outside an existing frame. It offers seven aspect-ratio presets, including square, wide, and tall layouts, or custom dimensions between 256 and 1536 pixels per side. Each generation returns one result for 10 credits, with JPG, PNG, or WebP selected before generation. Supported canvas size and download format do not change that published 10-credit price.
The 80s photo tool is a preset transformation rather than an open-ended canvas control. The vendor describes preserving a person's face, skin tone, and age while restyling hair, clothing, and surroundings with mid-1980s styling, analog grain, faded color, direct flash, and a date stamp. Identity preservation is a vendor claim, rather than an independently measured accuracy result.
That portrait tool accepts JPG, PNG, or WebP up to 10MB, costs 10 credits per photo, and advertises a 1024px JPEG download.
The video workspace supports creating from text, animating an image, using reference images, or editing an existing video. Available inputs and settings follow the selected mode. Models also differ in supported duration, resolution, sound, and reference options, so a control shown for one model is not a platform-wide entitlement.
The video credit price updates with the selected settings. Completed videos can be previewed and downloaded as MP4 within 30 days.
The A2E-oriented landing page includes an important boundary around its example gallery: the clips are provider demonstrations hosted for playback, rather than A2E outputs or ImageLayered test results. Previewing, copying a prompt, or filling recorded settings is free; clicking Generate creates a paid task. Unrecorded settings use the current model defaults.
The Qwen workspace exposes a specific way to distinguish multiple images and combine them with optional style controls:
<Picture 1> through <Picture 10> so the requested roles can be tied to individual inputs.Precision's people-extraction example sets the output count first, then adds a description of what to isolate. In the three-output example, the count includes one base image and two extracted layers. The optional prompt never changes the selected count, even when it asks for a layer for every person. Count and grouping instructions therefore serve different roles.
The downloadable assets are transparent PNG layers for moving, recoloring, or compositing in Figma, Photoshop, or another editor that supports alpha channels. The vendor describes the output as a stack of RGBA PNGs that approximates a layered project, with supporting elements separated when possible and hidden regions filled. That description preserves an important boundary: the result is a set of generated image layers, rather than a promise to recover the original authoring file.
The vendor positions layer separation for catalog photos where a product, props, and backdrop need to become distinct assets. Its examples describe changing colors, swapping backgrounds, or creating A/B variants from an existing photo.
Qwen's published campaign use case combines a requested layout, headline, and visual style with portrait, square, or landscape ratios for different channels.
The extender's wide example is positioned for a video thumbnail, presentation, or website banner. Added horizontal scenery creates room around the source composition and can leave space for a short title.
The extender says it cannot recover the true contents of areas that were never captured in the original image.
For separating a flat image into layers, Image2Layer is another option. Its site lists a layered PSD, transparent PNG layers, or a ZIP. It says users can choose 2 to 16 layers per run, while Auto selects up to 10.
ImageLayered describes transparent PNG layer downloads, and its Precision interface lists 2–13 outputs; its examples count the base image among those outputs. The two sites do not establish that their layer counts group the same image in the same way. Compare sample exports for your own editing workflow. This comparison concerns layer separation only: ImageLayered also offers a broader collection of image and video tools.
ImageLayered's Qwen generator presents four generation workflows in a browser workspace. The GPT and Nano Banana generator descriptions also advertise use without an API key or setup.
Availability after generation is described at tool level. The GPT generator says its PNG, JPG, or WebP results are saved to Recent generations for 30 days. Qwen likewise states a 30-day Recent generations period. The video workspace separately gives an MP4 download window of 30 days.
The general About text says that most generators do not save history. That generalization does not name the remaining tools it covers; the 30-day statements above apply to the named generators.
The public pricing capture on November 12, 2026 displayed the following monthly subscription cards. Amounts preserve the site's $ symbol; this capture did not identify a currency code. The visible monthly state is the basis for the table, without deriving annual prices from the advertised savings label.
| Plan | Displayed monthly price | Credits |
|---|---|---|
| Basic | $24.90 per month | 4,500 |
| Pro | $69.90 per month | 13,000 |
| Studio | $129 per month | 24,500 |
The subscription cards say that subscription credits expire at period end; subscriptions renew automatically and can be cancelled anytime. Separately, the terms' payments-and-refunds section discusses purchased credits and states, “Credits do not expire, but are non-transferable.” That sentence does not explicitly label the credit type; the subscription cards explicitly name subscription credits.
Purchased credits are generally non-refundable under the terms, except where required by law or expressly stated by the service. That purchase-refund condition is separate from refunds for a failed generation task.
Quick Layers lists 10 credits per output layer, while Precision Layers lists 20 credits per output layer. The pricing examples show four Quick layers using 40 credits and six Precision layers using 120.
The GPT Image 2.5 price example starts at 4 credits for text-to-image at 1K low quality. Costs for other settings are shown before generation. Background removal lists 10 credits per image, the extender 10 credits per generated extension, and the 80s portrait tool 10 credits per photo. Video prices update with settings instead of using that same fixed-image schedule.
The extender page says new accounts receive 10 welcome credits, enough for one extension at its 10-credit rate. It also says free generations enter a queue while paid accounts skip it. Its refund wording distinguishes a failed submission from failed or timed-out tasks that are refunded when the page checks the result. These are the extender's published conditions.
Users retain rights to content they submit and must have the rights, licenses, and permissions needed to submit and use it. The terms grant ImageLayered a limited, non-exclusive license to process, transmit, and temporarily store submitted images, prompts, and related outputs for requested features, provider processing, service protection, investigating reports, and legal obligations. The terms say this does not transfer ownership of the user's content.
That submitted-content provision is not a blanket statement of commercial rights or ownership for every generated output. The reviewed terms and privacy policy did not supply an explicit universal commercial-use grant or an explicit no-training commitment. Training use is not explained in those reviewed texts; silence does not establish either permission or a prohibition.
Uploaded images may be stored in private Cloudflare R2 object storage and made available to processing services through time-limited signed links. The privacy policy says there is no uniform immediate-deletion period across all tools and does not promise deletion as soon as a task finishes.
Necessary images, prompts, settings, and task data are sent to the selected AI provider. That provider's terms and policies govern its handling and retention of received data. ImageLayered's own policy separately describes retaining relevant information for service delivery, task status, billing or refunds, security, support, and legal obligations, with handling varying across tools.
Deletion can be requested, but the policy qualifies this with reasonable practicability and possible retention for law, accounting, security, dispute resolution, or provider policies.
The homepage also advertises removing text and logo watermarks from images and videos. Removing a watermark does not itself grant rights to reuse the underlying material; obtain the necessary permissions before editing or using third-party work. The content rules prohibit pornography or NSFW content, including explicit sexual acts, sexually intended nudity, and non-consensual intimate imagery. They also prohibit unauthorized or deceptive impersonation and deepfakes, including falsely attributed identity, statements, approval, or endorsement. Personalized-photo and video tools remain subject to those stated content boundaries.