Meet the Social Diffusion Generation Assistant
Meet the Social Diffusion Generation Assistant
Loading media image #1685...
Creating the image you have in mind shouldn’t require writing the perfect prompt on your first try.
The SocialDiff Generation Assistant gives you a conversational way to create, edit, improve, and explore images directly inside SocialDiff.
Tell the Assistant what you want in everyday language. It can turn a simple idea into a detailed generation prompt, transform an existing image, improve image quality with AI upscaling, or help you figure out the right style, lighting, composition, and creative direction.
One Assistant for the Entire Creative Process
You don’t have to decide whether you need a prompt generator, image editor, upscaler, or creative assistant before you start.
Just describe what you want.
The Generation Assistant understands your request and determines the appropriate action.
You can use it to:
- Create a completely new image from text
- Transform or restyle an existing image
- Replace, remove, or add elements
- Change lighting and atmosphere
- Reframe an image or change its orientation
- Upscale and enhance image quality
- Develop better prompts
- Get advice about styles, composition, cameras, lighting, and aspect ratios
Here’s what that looks like in practice.
1. Turn a Simple Idea Into an Image
You don’t need to write a long, technical AI prompt.
Start with the idea.
Try asking
“Create a lonely lighthouse on a cliff during a violent storm.”
The Assistant can expand that concept into a much richer visual direction, considering things such as composition, environment, atmosphere, lighting, materials, camera treatment, depth, color, and finishing details.
Your short request might become an image with dramatic waves crashing against dark rocks, rain cutting through the scene, a powerful lighthouse beam illuminating the mist, cinematic contrast, and carefully composed negative space.
Loading media image #1663...
You stay focused on the idea while the Assistant handles much of the prompt engineering behind it.
More Generation Ideas
Cinematic Portrait
“Create a cinematic portrait of an astronaut inside a greenhouse.”
Loading media image #1664...
Product Photography
“Create a premium studio photograph of a futuristic silver sneaker floating above black reflective glass.”
Loading media image #1665...
Fantasy Environment
“An ancient library hidden inside a gigantic hollow tree, warm magical light, tiny figures exploring it.”
Loading media image #1666...
Editorial Fashion
“Create a high-fashion editorial portrait with brutalist architecture, overcast light, and a restrained monochrome palette.”
Loading media image #1667...
2. Edit Images by Simply Describing the Change
Already have an image that’s close to what you want?
Attach it and tell the Assistant what should change.
Instead of rebuilding your image from scratch, the Assistant can use the existing image as the visual starting point and perform creative transformations.
For example:
“Transform this into an elegant Art Nouveau poster.”
Loading media image #1669...
The Assistant is designed to understand the change you’re requesting while keeping compatible parts of the original image intact.
3. Change the Style
Sometimes the composition is already right—you just want an entirely different visual language.
Attach your image and try:
“Turn this into a watercolor storybook illustration.”
Or even simply:
“Watercolor storybook.”
The Assistant understands short style instructions attached to an image as editing requests.
Loading media image #1671...
Other Ideas You Could Try
“Restyle this as a hand-painted fantasy illustration.”
“Turn this into a retro 1980s sci-fi movie poster.”
“Make this look like a luxury fashion campaign.”
“Transform this into a dark gothic oil painting.”
“Give this a soft clay-animation look.”
4. Replace Objects and Visual Elements
The Assistant isn’t limited to applying filters.
You can describe specific changes to what’s inside an image.
For example:
“Make her wear a leather jacket.”
Loading media image #1672...
You can make requests such as:
“Remove the people in the background.”
“Replace the cloudy sky with a star-filled night sky.”
“Add wildflowers throughout the foreground.”
“Change the blue jacket to a black leather jacket.”
“Replace the modern buildings with medieval architecture.”
The attached image remains the source of truth, so your instruction can focus primarily on what needs to change.
5. Completely Change the Mood With Lighting
Lighting can transform an image without changing its fundamental subject.
For example, upload a daytime image and ask:
“Turn this into a rainy neon-lit night scene.”
Loading media image #1674...
Try Variations Such As
“Make this golden hour with warm sunlight coming from the left.”
“Give this dramatic studio lighting with strong rim light.”
“Make the scene feel like an early foggy morning.”
“Add soft window light and a premium editorial mood.”
This makes the Assistant useful for exploring several visual directions without rebuilding the entire composition.
6. Turn Portraits Into Landscapes—and Landscapes Into Portraits
Need an image for a different format?
Ask the Assistant to change the canvas.
For example:
“Make this a cinematic horizontal scene at sunset.”
The Assistant can expand or reframe the scene rather than simply stretching the existing image.
Loading media image #1675...
You can ask for:
- Square images
- Portrait / vertical images
- Landscape / horizontal images
This is especially useful when adapting artwork for different placements while keeping the same visual concept.
7. Upscale and Restore Images With AI
Not every image needs a creative transformation.
Sometimes you simply want a better version of the image you already have.
Attach an image and ask:
“Upscale this image.”
Or be more descriptive:
“Restore and enhance the details.”
Loading media image #1676...
The Assistant recognizes when you’re asking for quality enhancement rather than a creative edit and uses the dedicated AI upscaling process.
Your original aspect ratio is preserved.
8. Ask for Creative Advice Before Generating Anything
The Assistant doesn’t have to create an image every time you send a message.
You can use it as a creative partner first.
For example:
“What lighting would work for a mysterious cinematic portrait?”
Or:
“How can I make a product photo feel more luxurious?”
Or:
“What aspect ratio should I use for a cinematic landscape?”
The Assistant can help with:
- Prompt writing
- Artistic styles
- Lighting
- Camera and lens direction
- Composition
- Framing
- Aspect ratios
- Image-generation concepts
- Image-editing approaches
Loading media image #1677... Loading media image #1680... Loading media image #1679...
This means you can explore an idea conversationally and then turn that direction into an image when you’re ready.
9. Continue Refining the Idea Through Conversation
Image creation rarely ends after one instruction.
The Assistant understands the conversation around your current creative direction, allowing you to develop an idea naturally.
For example:
You:
“I want a premium editorial look.”
Assistant:
Suggests a restrained palette, directional lighting, and a cleaner composition.
You:
“Make it more cinematic.”
You:
“Now make the scene horizontal and change it to sunset.”
Want to Continue From a Specific Generated Image?
Pin the image you want to work on and the Assistant will switch to it.
Loading media image #1682...
One idea, refined through conversation.
Instead of repeatedly reconstructing complex prompts, you can keep refining the result naturally through the same conversation.
10. Detailed Prompts Still Work
The Assistant isn’t only designed for beginners.
If you already know exactly what you want, give it a detailed brief.
For example:
“Create a vertical cinematic portrait of an explorer standing inside an enormous abandoned greenhouse. Keep the person in the lower third of the composition and emphasize the scale of the architecture. Morning fog should fill the background, with warm sunlight entering from broken glass panels on the upper left. Use muted greens, weathered brass, dark fabric, shallow atmospheric perspective, subtle film grain, and a premium editorial photographic treatment.”
The Assistant preserves explicit requirements instead of replacing your detailed direction with a generic prompt.
That includes requirements around composition, dimensions, preservation, style, relationships between subjects, and other visual constraints.
Loading media image #1683...
Give the Assistant a simple idea—or specify every detail.
A Few Things to Know
-
One source image at a time. The current Generation Assistant works with one source image at a time.
-
Choose your source image. The source image can be selected from your gallery using the Images button, uploaded directly, or selected by activating the Pin on an image generated in the current chat.
-
Whole-image editing. The Assistant currently focuses on whole-image conversational editing rather than painted masks or region-specific mask inputs.
-
Start a new generation. When you want to generate a completely new image, simply select New Chat.
-
Upscale your final result. When you’re happy with the result, ask the Assistant to upscale the image for a sharper, cleaner final version.
-
Pricing. Creating a new image with the Assistant costs 30 credits, editing or transforming an existing image costs 60 credits, and AI upscaling costs 20 credits.
What Will You Create?
The goal of the SocialDiff Generation Assistant is simple:
Spend less time figuring out how to communicate with an image model and more time deciding what you want to create.
Start with a sentence:
“Create a cinematic portrait…”
Attach an image:
“Turn this into a painting…”
Pin a generated image and change one detail:
“Make the jacket black…”
Explore a new direction:
“What would make this feel more premium?”
Or improve something you already created:
“Upscale this and restore the details.”
The Assistant handles the rest of the workflow from the same conversation.
Generate. Edit. Transform. Upscale. Just ask.
Try the Generation Assistant in SocialDiff
Open the Generation Assistant, choose an image or start from nothing, and describe what you want to create.