Briefly, it’s a sandbox with the best AI tools in one workspace available in your browser.

Phygital+ is a web node-based AI workspace for creators to work with neural networks without code.

Phygital+ gives users the ability to work with several neural networks and tools at once and in one space, making the content creation simpler and more creative.

Plus, with the node system you can combine multiple AI tools into one workflow, save it and share it with anyone!

Node is a block that you can move on the workspace. Each node corresponds to a certain neural network, feature, or task.

Each node has one or a few parameters that you can change in order to get different results.

Connecting nodes with each other through 'input' and 'output' allows you to create a chain of different tasks.

For quicker navigation and seamless workflow you can copy-paste nodes by using hotkeys Ctrl+C and Ctrl+V

AI Library is a catalog of AI tools and neural networks with filters and semantic search. It’s a database for Generative AI newest and latest developments, an ultimate aggregator of information for Generative AI.

AI Library is created and curated by Phygital+ team. The most popular tools from AI Library are added to Phygital+ node interface. Also AI Library is aimed at educating people on how to use AI, that’s why there’s a blog page and page with tutorials for the most common AI models.

In our product at the top panel you can find ‘Learn’ section with a Quick Tour and Tutorials page. We highly recommend taking a Quick Tour in the app, checking out Quick start page and watching the Onboarding into Phygital+, as these materials show the best the possibilities of Phygital+ product.

You can also click on the ? icon in the top bar to see the main information about the application.

For more additional tutorials and information, refer to our Docs page with Tutorials, Articles and best use cases.

Certainly! We have in-depth tutorials to the most used AI tools, which you can find at our Docs in the AI Tools (Nodes) section. For more tailored guides for specific use cases refer to the tutorials page with recent guides and articles.

No, if you close the tab, the generations won’t be saved. However, before you close the page you can click on Save project in order to save the nodes with its results.

The training will continue if you close the tab. After it’s done, the new LoRA will appear in My loras in Stable Diffusion XL.

We have a special offer for creative teams which include exclusive discount, private educational sessions, private support, unlimited generations, Access to all tools, Teams templates and Collaborative features.

You can request personal onboarding by submitting the form, contacting our product manager Olga in Telegram or by booking a call.

Phygital+ is a sandbox with many awesome AI tools on one canvas. On this canvas we have little blocks called nodes, which you can connect to each other. We like to compare our nodes to LEGO blocks, as with the nodes you can build a workflow – for example, creating text from image, removing background on it and upscaling it.

We highly recommend taking a Quick Tour to get the basics of our interface and we are sure you will get the hang of it! :)

If you have a VPN program installed on your PC (including educational, corporate, or work-related VPNs), it may interfere with node operations, even when inactive. Close or disable the VPN completely and restart your workflow to resolve the issue.

1. Add the node on the workspace. You can do it by clicking on any AI tool in left side panel.

In Phygital+ you can create images using the most popular AI tools:

• Stable Diffusion XL,
• FLUX
• DALL-E 3,

*For creating images from text we recommend using Stable Diffusion XL, as it creates images in high quality.

2.Type any text in the big Text prompt field
3.Press Start button.

We recommend trying two AI tools: Stable Diffusion XL or DALL-E 3.

With Stable Diffusion you will more likely get a photorealistic concept, and DALL-E 3 creates more digital looking images.
DALL-E 3 gives very good results with short prompts, whereas Stable Diffusion XL requires a more thorough work with prompt.

New Model Flux also generates incredibly detailed images and follows the prompt as well as DALL-E 3 does.

You can ask the AI Assistant in the lower right corner or hover over the parameter in the node and a short description of it will appear on your screen. Also in Docs you can find the detailed instruction for the most features.

You can perform quick actions with any image by hovering over it:

• Preview – opens up a full-screen preview of the image
• Save – saves the image to your PC / laptop
• Upscale – upscales the image by creating a node, automatically connecting it and launching it
• Inpaint – adds image to the inpainting window in the right side of the screen
• Variations – creates similar images in SD XL.

The best model for generating images is Stable Diffusion XL (SD XL). It can generate anything in high quality (1024x1024).
New Stable Diffusion 3 is more preferable to use for generating landscapes, objects.

*The advised Models to use in SD XL are Juggernaut 8, RealVis 4 or RealCartoon

A model (style) in Stable Diffusion refers to a particular model that has been trained on some style, object or person.
The most popular models are those that have been fine-tuned (improved) to generate highly photorealistic results with fewer artifacts.
By using a custom style you are almost guaranteed to get good results with short prompts.

*The best options for Stable Diffusion XL: Juggernaut, RealVis 4, RealCartoon

You can do it in four ways:

1. Hover over your generation and press Variations

2. Use Stable Diffusion XL img2img
– Add the Stable Diffusion XL (img2img) node to the workspace
– Connect your image as Start image
– Copy the prompt from your first node (or get a new one using GPT-4o node with prompt 'Make a prompt for Stable Diffusion'
– Press Start
– If you want your image to change more, set Start Image Skip (Denoising strength) parameter to 0.6 or 0.7

3. Use IP Adapter
– Add IP-Adapter XL node
– Connect your image as Style (IP image)
– Press Start

4. Using Describe Image
– Upload your image to the workspace
– Add Describe Image node and launch it
– Create SD XL node, past your prompt from Describe Image and press Start"

You can do it using IP-Adapter XL.
Do the following:
1. Add your style to the workspace
2. Create IP-Adapter node
3. Connect your style reference to Style (IP image)
4. Connect the image which you want to change
5. Turn on Use ControlNet in the settings and Type Canny
6. Press Start

Option 1.
Using AI Training with LoRA Train. Then you can generate images using Stable Diffusion XL (ControlNet XL, IP-Adapter XL)

Option 2.
Recommended: Using IP-Adapter. Do the following:
1. Add your style to the workspace
2. Create IP-Adapter node
3. Connect your style reference to Style (IP Image)
4. Press Start to create similar images
5. You can also type any prompt to create images with your idea in this style
*If you want your style to affect the image more, increase IP Adapter Weight.

1. Add your image to the workspace (by drag-and-dropping it or copy-pasting it to the workspace)
2. Hover over the image and press Upscale

You can also use Creative Upscale, which adds more details to your image and can be used for restoring old photos

If you generated image in Stable Diffusion XL, you can increase detailisation by locking the seed (press the Lock icon next to the Seed parameter), choose LoRA (Modifier) Detail Enhancer or Detail Tweaker and restart the node)

1. Add your image to the node workspace (by drag-and-dropping it or copy-pasting it to the workspace)

2. For removing background: add Remove background node

3. For replacing part of the image:
– Hover over the image and press Inpaint
– In the right side panel you will see your image, now you should draw over the part of the image you want to replace.
– In the prompt below the image type what you want to see instead of the object on your image
– Press Generate
– If you don't like the result, press Start again, change Seed or Models (we recommend trying Juggernaut, Deliberate 5 or Reliberate 3)

For increasing resolution and enhancing quality we recommend using Upscale image node.

If you want to remove blur from the image, try using Stable Diffusion XL (img2img) for it.
1. Create SD XL (img2img) node
2. Connect your image
3. Set Start Image Skip (Denoising strength) to 0.1
4. Write down the prompt
- copy the prompt from your original generation
- or get the prompt using GPT-4o (use the prompt ‘Make prompt for Stable Diffusion 3’)
5. Press Start

Check if your seed is locked in the Settings of the node, if there's a checkbox marked, turn it off

If your generation looks resized or shrinked while generating, make sure that you set the same aspect ratio and resolution in the Settings of the node.
IP-Adapter XL and ControlNet XL have the special settings for setting resolution of the final generation, meaning it will not take the resolution from the original image. Therefore, if you connect an image with 512x512 resolution, the generation will be in 1024x1024, and nothing will look weird.

However, if you connect the image with another aspect ratio (like 16:9), the node will resize it to 1024x1024, which will cause the image, and the generation will look bad.
Therefore our tip is to set resolution manually in the settings if you do not use square images (any image on the workspace has the resolution written below it).

Try using the Face Restore node. Simply connect your image to it and press Start.
If you’re using Face Swap, we recommend switching to Face Enhancer (and choose model GAN), as it can improve general facial features without changing the original person too much.

Option 1.
The easiest way: You can try the node Automatic Image Mask – connect your image, type in the prompt what you want to select and press Start. It will create the mask automatically for you

Option 2.
Recommended if you want to have more freedom while creating the mask.
Hover over your image and press Inpaint. In the window to your right you can draw over the area which you want to select. After you’re done, click ‘Mask’.

*To change an already created mask, hover over the image, in the window to your right draw over the area you want to change with black or white. Press ‘Image + Mask’"

1. In our product you can generate a 3D model of your object using 3D From Image node
2. Also you can generate textures for 3D by using Midjourney node and turning on Tiling parameter in the settings.
3. By using ControlNet XL you can transform UV texture into a completely different one
4. Quickly preview textures or UV maps on your 3D model using 3D Object node.

Currently we have a new tool called 3D from Image that can create a 3D object from your image. You can use it for creating different views of your object.
1. Connect your image to 3D from Image and press Start to get a 3D model
2. Zoom in to your model in 3D from Image node, turn it how you want and press Render
3. Create IP-Adapter XL node, connect your 3D render to Start Image, and your original image with object as Style (IP image). Add the prompt (or create it using Describe image) and select Use ControlNet in the settings. Press Start to get your result

1. In Phygital+ you can create a video with quick interpolation between frames from several images using Morph video effect node.
2. You can create a quick 4 second video from a simple image using Stable Diffusion Video node.

LoRA is a modifier which is aimed at improving generations in Stable Diffusion XL nodes.
LoRAs are available for Stable Diffusion XL, Stable Diffusion XL (img2img), ControlNet XL, Inpainting XL, IP-Adapter XL.
LoRAs can increase detalization, remove artifacts, make images more artistic or help you achieve specific style.
If you want to increase detalization, use LoRAs Detail Tweaker or Detail Enhancer. If you want to achieve a specific style, try LoRAs like ‘Style: Sticker’, ‘Style: Glowneon’. We highly recommend trying LoRA ‘Midjourney Style’

In order to use LoRA, follow these steps:
1. Create any Stable Diffusion XL node
2. Type your prompt
3. Go to the Settings and find Modifiers. Use up to 3 modifiers at the same time.
4. Press Start
Additionally, you can train your own LoRA to use it in any generation in Stable Diffusion XL.

Some models in Stable Diffusion XL do not generate good results with some LoRAs. If you get artifacts using LoRA, try changing the Model or adding the negative prompt like ‘blurry, ugly, weird’

Midjourney is now available only for Teams and for Plus users, who have subscribed before June 2024. New Plus subscribers don’t get the access to Midjourney.
You can still get access to Midjourney if you subscribe to Teams Plan (available for 1 Member as well).
We are offering a special team plan that includes access to Midjourney, please contact us filling this form.

Pro Tip: For mimicking Midjourney Style you can use LoRA ‘Midjourney Style’. You can find it in any Stable Diffusion XL node.

deformed hands, worst quality, low quality, illustration, 3d, 2d, painting, cartoons, sketch, distorted hands, extra fingers, extra hands, deformed face

You can submit a request for a new feature by submitting the form

Currently the only person who can see your images is you. So, you don’t have to worry about your privacy. However, if you create images violating legal and ethical norms, our system will detect it and you will get a warning. Further repeated illegal actions after warning will result in banning your account.
Learn more about Privacy here

You can use the images for your commercial projects, if you are a paid subscriber to Phygital+.

We can’t guarantee the full and reserved rights for your user generated content, as you might be using the data which is not owned by you.

By using Phygital+ you represent and warrant that neither your User Assets, nor your use and provision of your User Assets to be made available through our Services, nor any use of your User Assets by Scenario on or through our Services, will infringe, misappropriate or violate a third party’s intellectual property rights, or rights of publicity or privacy, or result in the violation of any applicable law or regulation.

Read more in our Terms and conditions.

In the workspace click on the circle in the top right corner, and there you can find an option to change password.

Usually sending the email one more time helps. We also recommend to make sure that you signed up using an email and a password, and not Google Authorization. In case of Google Authorization, you don’t have a password, simply login using Google again.

AI Library is completely free-to-use.
AI Workspace has Freemium option, all you need to do is to sign up. However, you get limited access to some nodes and have a limit of generations of 10 per week.

More about pricing plans for AI workspace here

The quickest way is to press on ‘Get premium access’ banner on the top right corner or on the pricing page

Click on the top right corner on your avatar, go to your account page and there you have to press on Manage subscription. You will see the banner with all available plans, where you can find the button Downgrade.

Currently we are offering Freemium access to almost all features with the limit of 40 generations per account per month or 10 generations per week.

As a free user for your first 10 tasks you get access to all AI tools.
Afterwards you have access to 11 AI Tools with which you can generate, transform and enhance images.
You have access to: Stable Diffusion XL, Stable Diffusion XL (img2img), Inpainting, Inpainting XL, Remove Background, Face Enhancer, Upscale Image, Face Restore, Describe Image, Prompt Extend, Depth Map from Image.

As an advanced user you get access to 13 tools for generating, transforming and enhancing images.
You also get a higher limit for generation with 500 runs.

You have access to: Stable Diffusion XL, Stable Diffusion XL (img2img), Inpainting, Inpainting XL, Remove Background, Style Transfer, Face Enhancer, Upscale Image, Face Restore, Describe Image, Prompt Extend, Depth Map from Image, Automatic Image Mask.

As a premium user you get access to 24 tools for generating, transforming and enhancing images, and AI training.
You get limited access to generations and 10 AI trainings

You have access to: AI Training with LoRA Train, Stable Diffusion XL, Stable Diffusion XL (img2img), DALL-E 3, IP-Adapter XL, Stable Diffusion Video, ControlNet XL, Change Light, Inpainting, Inpainting XL, Morph Video effect, Remove Background, Style Transfer, Face Swap, Face Enhancer, Upscale Image, Face Restore, GPT-4o, Describe Image, Prompt Generator, Prompt Extend, Depth Map from Image, Automatic Image Mask, 3D from Image.

As a plus user you get access to all 27 tools for generating, transforming and enhancing images, and AI training.
You get unlimited access to generations.

You have access to: AI Training with LoRA Train, FLUX, Stable Diffusion 3, Stable Diffusion XL, Stable Diffusion XL (img2img), DALL-E 3, IP-Adapter XL, Stable Diffusion Video, ControlNet XL, Change Light, Inpainting, Inpainting XL, Morph Video effect, Remove Background, Style Transfer, Face Swap, Face Enhancer, Creative Upscale, Upscale Image, Face Restore, GPT-4o, Describe Image, Prompt Generator, Prompt Extend, Depth Map from Image, Automatic Image Mask, 3D from Image.

• Special pricing offer! Exclusive Discount
• Private Educational sessions
• Private Support
• Unlimited generations & GPU Time
• Unlimited styles & AI tools
• Access to Midjourney, FLUX, DALL-E 3 and Stable Diffusion 3
• Teams Templates and Collaborative Features

Contact us through the form.

If you have subscribed before June 2024 to Plus, your access to Midjourney includes 50 fast generations per day, making it up to 19h fast hours (vs 15h fast hours in 30$ Basic plan of Midjourney). After the fast generations per day you are still able to generate images in relax mode, it's switched automatically. Number of total Midjourney available generations is 2000 per month.

With DALL-E 3 you get 100 generations per day if you’re a Plus subscriber, and 100 generations per month if you have Premium subscription.
Stable Diffusion Video limits are the following:

• For Premium users - 20 per month.
• For Plus users - 50 per month

*GPU Time refers to the total available generations. Each pricing plan has a limit for the overall generations number per billing month (for example, Advanced has 500 limit of generations and Premium – 1000 per month). Upon reaching this number the daily limit is turned on. All the following days until the end of billing period the user keeps the ability to use all perks of their plan, but with these restrictions:

• Advanced have 50 generations that reset daily.
• Premium users get 100 generations that reset daily.

**The trained models are saved on the Phygital+ servers for 6 months, afterwards the team reserves the right to delete them if they are not used by the user within this period. The subscribed users can download their models beforehand through the Train Panel interface.

Currently we don't have open Free Trial, however, in our Teams offer you can get Free trial for your creative team. Contact us through the form

We are offering an alternative method of payment via other cards. Switch to 'Sber / Mir' on the Pricing page in the app.

We highly recommend you to subscribe to our best plan ‘Plus’ as it includes access to all available tools which you might ever need to work with AI.

Email us at ai@phygital.plus, fill in this Support form or write in our Discord server for any inquires, requests, suggestions and questions.

According to the Section 3.2 of the Terms and Conditions, you can subscribe to the Phygital+ services on a regular (monthly or 6-monthly) basis according to the prices and terms on the pricing page, and we will charge your payment method on an agreed-upon periodic basis unless you cancel your subscription on the account page. All payments are nonrefundable except the situations where a technical error occurred with the payment (e.g, you have been charged twice due to the system errors).

If you have subscribed to Phygital+ using Stripe, write to us at ai@phygital.plus.

With the alternative method (such as MIR or Sber), you get the receipt right after you subscribe to Phygital+.

With AI training you can enhance your business workflows and perform certain creative tasks easier.
AI Training is the process of adding the information about any subject (person, object, style) in the memory of the neural network (Stable Diffusion XL). This process takes around 20 minutes and after the training is complete, you can start creating any images with the trained subject in Stable Diffusion XL. This gives opportunity to do the following:
· create marketing materials with any celebrity,
· create character concepts with your character reference,
· create SMM content with your product,
· create AI avatars in different style with any person,
· create branded materials and design assets with your unique style.

In order to train AI you have to be at least a premium user to unlock AI training with LoRA Train node
1. Prepare your images on which you will train.
2. Upload images to the workspace using Import Files node, or drag and drop them
3. Connect all images to LoRA Train node, write your unique name (it should be something unknown to Stable Diffusion or generally the Internet. For example, DariaWnd. Add to the end of your unique name ‘1woman’, ‘1man’ - basically 1 and the type on which you train
4. Set the parameters: usually the default settings are good for training on 6-7 photos. If you train on 13-15, you can increase the number of steps to 2500
Use Juggernaut 8 or RealVis 4 model for training.
Optionally: set Learning rate to 0.0001 to get higher detalization
5. Press Start and wait for the result. Usually training takes up to 40-50 minutes

You can also watch our tutorials for reference

1. Prepare your images on which you will train.
2. Upload images to the workspace using Import Files node, or drag and drop them
3. Connect all images to LoRA Train node, write your unique name (it should be something unknown to Stable Diffusion or generally the Internet. For example, Gucdres. Add to the end of your unique name ‘1dress’, ‘1bag’ - basically 1 and the type on which you train
4. Set the parameters: usually the default settings are good for training on 6-7 photos. If you train on 13-15, you can increase the number of steps to 2500
Use Juggernaut 8 or RealVis 4 model for training.
Optionally: set Learning rate to 0.0001 to get higher detalization
5. Press Start and wait for the result. Usually training takes up to 40-50 minutes

Watch the tutorial here

1. Prepare your images on which you will train.
2. Upload images to the workspace using Import Files node, or drag and drop them
3. Connect all images to LoRA Train node, write your unique name (it should be something unknown to Stable Diffusion or generally the Internet. For example, HenrkAarrestad.
4. Set the parameters: usually the default settings are good for training on 6-7 photos.But if you are training on style, we recommend using at least 25 photos. Increase Training steps to 2500 at least
Use Juggernaut 8 or RealVis 4 model for training. Optionally: set Learning rate to 0.0001 to get higher detalization
Set Dataset Repeats to 20
5. Press Start and wait for the result. Usually training takes up to 50-60 minutes

After the training is complete, you can use LoRA in any Stable Diffusion XL node: Stable Diffusion XL, Stable Diffusion XL img2img, ControlNet XL and IP-Adapter XL.
It will be added to the list ‘My loras’. By selecting your trained LoRA, the right keyword will be added to your prompt.

If you have successfully trained AI model, your trained LoRA will appear in the list ‘My Loras’ in Stable Diffusion XL, Stable Diffusion XL (img2img), ControlNet XL and IP-Adapter XL