AI for Photoshop that works on your layers.
Sekondo connects Claude or ChatGPT to Photoshop. You tell it what you want and your assistant works in your file, with your layers. No flat image: everything stays editable.
Adobe’s AI, just by asking.
It works with Adobe’s latest, beta versions included. You give a command; it decides which tool to use at each step.
In your own words you ask for, say, a perfume ad, and Sekondo does the rest: it writes every prompt in full detail and chains the tools together on its own, one after another. It generates the bottle with Magnific, removes the background and creates the setting with Firefly, harmonizes the light, adds water drops to the glass with Nano Banana and upscales with Topaz. It picks the model that fits each step best (Firefly, GPT Image 2, Nano Banana or FLUX) and leaves everything in layers, editable.
Sekondo already knows you when you ask for a complete poster in a single request. It gives it to you with smart, editable layers.
“Make me a perfume poster in this blank document, in layers, following our client’s corporate identity and these directions:
An amber bottle with a gold cap and no label, made with Magnific, on a wet rock at the shoreline at sunrise, with the light harmonized; then, on another layer, water drops on the glass.
On the bottle, our Sekondo SK badge in gold with reflections.
At the top, centered, AMANECER and below it EAU DE PARFUM.”
What Sekondo does in Photoshop.
It learns from you, and this is just a sample. If Photoshop can do it, Sekondo does it.
- Runs Photoshop’s AI just by asking: Generative Fill in the area you name, removing and generating backgrounds with their shadow and reflection, and harmonizing the light so everything fits.
- Generates images with the model you name: Firefly, GPT Image 2, Gemini 3 (Nano Banana Pro) or FLUX.2 pro.
- Edits the photo from a description, “put snow on the rocks”, and leaves the rest as it was.
- Doubles the size of your images with Topaz generative upscaling.
- Builds a whole ad from a single request, from blank canvas to headline, every layer with its name.
- If your assistant has another image AI connected, it uses it alongside Photoshop’s in the same request.
- Learns your retouch on the first photo and repeats it across the whole batch, measuring every photo before touching it.
- Cuts out the product, places it in your template and adds the shadow, the reflection and the gradient the way you do.
- Sends the image through generative AI with a prompt written for each photo, not a generic one, and brings it back to its own layer.
- Adjusts each photo in a batch separately: backlighting, mixed light, subject and background each in its own way and on separate layers.
- Runs your saved actions and scripts and combines them, with rules on which layers or files they touch.
- Recomposes a design in another format, from landscape to square for example, repositioning every layer.
- Swaps the photos inside your templates’ smart objects, each one fitted in place, without rebuilding any.
- Translates posters without losing the typeface or the style of each word, and if you ask, duplicates the text layers by language, sorted into groups.
- Organises, renames, shows or hides layers with one command.
- Finds backgrounds and textures for your composition.
- Checks the spelling of every text layer.
- Exports at the resolutions you need - catalogue, web or social - with their colour profile. If you ask, with the name you want; on its own it does not rename or delete any file.
- Removes the background from every layer or image in a batch and gives them a colour background.
- Creates a campaign’s image sets with their variants of size, composition and text, and exports them in one go.
- Adds a watermark to every image, with the position, opacity and size you tell it.
- Writes the alt text for the images.
- Exports layers by their name, their visibility or the condition you set, to several folders.
- Builds adjustment layers with their masks and blend modes, for example Curves with Color Balance.
- Adjusts the exposure, colour, highlights and shadows, hue and saturation of each image.
- Selects the subject, blurs the background and applies Lens Blur.
- Extends the canvas when the image is short of space.
- Adds grain, noise, halftone or a colour tint.
How to brief it.
Like someone who just walked into the studio: it knows Photoshop and knows nothing about your client. You tell it which document, what you want and the judgement to apply. Before touching anything it reads the whole thing, and when it is done it tells you what has changed. Each command is a single step in the history: if you do not like it, Cmd+Z and it is back as it was.
It works on your computer, with your Claude or your ChatGPT. The file never leaves your machine.
Photoshop tasks Sekondo solves, step by step.
How do I rename many layers in Photoshop at once?
Sekondo organizes, renames, shows or hides layers with one order.
How to ask: “In Poster.psd, rename the layers after what they contain and hide the ones that aren’t used.”
How do I apply the same retouch to a whole folder of photos in Photoshop?
Sekondo learns your retouch on the first photo and repeats it across the whole batch, measuring each photo before touching it. It adjusts each photo in a batch separately: backlight, mixed light, the subject and the background, each in its own way and on separate layers.
How to ask: “Learn the retouch from the first photo and apply it to the rest of the folder, measuring each one.”
How do I change the photo in a smart object template without redoing anything?
Sekondo swaps the photos inside your templates’ smart objects, each one fitted in its place, without redoing any.
How to ask: “In the Catalogue.psd template, replace the photos in the smart objects with the ones in this folder.”
How do I export an image to several sizes and color profiles?
Sekondo exports at the resolutions you need (catalog, web or social media), with their color profile; on its own it never renames or deletes any file.
How to ask: “Export this image for catalog, web and social media with its color profile.”
How do I make a product ad with Photoshop’s AI in a single request?
Sekondo runs Photoshop’s AI step by step in your document: it generates the product, removes its background, puts it in a setting with its shadow and reflection, harmonizes the light, adds the headline and upscales the result. Every layer with its name and everything editable.
How to ask: “Make an ad for this perfume on a rock at sunset, with drops on the bottle and the headline AMANECER.”
Works on your computer, with your Claude or your ChatGPT. The file never leaves your machine.
Frequently asked questions about Sekondo in Photoshop.
Can the AI change a photo for me in Photoshop?
Yes. With Sekondo, your Claude or your ChatGPT works inside Photoshop: it adjusts the exposure, colour, highlights and shadows, changes the background or edits the photo from a description. Each change stays on its own layer, editable.
How to ask: “Lighten the shadows on the face and change the background to a beach at sunset.”
Can it remove the background of a photo?
Yes. It removes it from one photo or from a whole batch, and if you want it generates a new one with its shadow and reflection.
How to ask: “Remove the background from all the photos in this folder and give them a white background.”
Can it use Photoshop’s Generative Fill?
Yes. It runs it in the area you name, generates images with the model you prefer (Firefly, GPT Image 2, Nano Banana or FLUX) and harmonizes the light so everything fits.
How to ask: “Fill the selected area with a vase of dried flowers and harmonize the light.”
Can it retouch a hundred photos at once?
Yes. It learns your retouch on the first one and repeats it across the whole batch, measuring every photo before touching it: backlighting, mixed light, the subject and the background, each in its own way.
How to ask: “Do to all the photos in the folder what I did to the first one.”
Can it make a whole ad for me?
Yes. From a blank canvas to the headline in a single request: the product, the setting, the light and the text, every layer with its name.
How to ask: “Make me an ad for this perfume on a wet rock at sunrise, with the headline at the top.”
Can it translate a poster without losing the typeface?
Yes. It translates each text keeping the typeface and the style of every word, and if you ask it duplicates the layers by language, sorted into groups.
How to ask: “Translate the poster into English and French and separate the languages into groups.”
Can it change the photos in my templates?
Yes. It swaps the photos inside the smart objects, each one fitted in its place, without redoing any.
How to ask: “Put these six photos in the template’s frames, in this order.”
Can it upscale an image without losing quality?
Yes. It doubles the size with Topaz generative upscaling.
How to ask: “Upscale this photo to twice the size so I can print it large.”
Can it use my actions and my scripts?
Yes. It runs your saved actions and scripts, combines them and runs them across a whole folder with the names the client asks for.
How to ask: “Run my ‘Web’ action on all the photos and save them with the client’s name.”
Does it work with Claude and with ChatGPT?
Both, in their desktop app and with your own subscription. The details are in the Docs.
What do I need to use it in Photoshop?
Photoshop 2024 or later, on macOS 11 or later or Windows 10 or later, and Claude or ChatGPT desktop with your subscription. The installer puts the panel in its app and adds Sekondo to your assistant: there are no commands to type.
You create.
Sekondo does the rest.
