What is WYSIWYG Web Builder AI?
AI has emerged as a pivotal tool in the quest for enhanced efficiency and productivity, playing a crucial role in reducing the time and effort required by creators in various tasks, spanning from content creation to coding. With the introduction of WYSIWYG Web Builder AI, the convergence of AI and website development has reached new heights, empowering users to construct and maintain websites in a mere fraction of the time and effort that was previously demanded.
WYSIWYG Web Builder AI boasts a robust set of capabilities, including:
- Text Generation: It can generate textual content.
- Text Enhancement: This tool can improve text by making it shorter, longer, simpler, correcting spelling errors, or adjusting its tone.
- Translation: WYSIWYG Web Builder AI can proficiently translate text from one language to another.
- Image Generation: It has the capacity to create images based on textual input.
- Image Manipulation: Users can generate variations of existing images and make modifications via in-painting or out-painting techniques.
- AI Vision: AI Vision can recognize what's in an image and lets you ask questions like, "What's in this image?" or "Write a caption for this image for use on a website".
- Background Removal: This AI automatically removes backgrounds from images.
Where can I learn more about the AI tools?
Please see these related tutorials:
Do I need a subscription?
The AI tools in WYSIWYG Web Builder are implemented using the OpenAI API. You will need an OpenAI API key in order to use this functionality.
If you do not already have an OpenAI account then you can create it here: https://platform.openai.com/signup
Note that this is not the same as a ChatGPT Plus account!
You can create an OpenAI API key for free. New users get free $5 worth of free tokens. These tokens expire after 3 months.
After the quota has passed you can choose to enter billing information to upgrade to a paid plan and continue your use of the API on pay-as-you-go basis.
- Pay as you go - You pay only what you use. For example, if you generate one 512x512 image, then the cost will be $0.018. Hundred (100) 1024x124 images will cost approximately $2.00 (2 dollar).
- Prepaid credits - You can also buy a prepaid bundles. For example $5.00 or $10.00.
Create an OpenAI account: https://platform.openai.com/signup
Create an API key in the account: https://platform.openai.com/api-keys
The API key can be entered in Tools -> Options -> Artificial Intelligence
Setup the billing information: https://platform.openai.com/account/billing/overview
Here you can buy credits or 'pay-as-you-go'
Here's an example bill to give you an idea of the costs:
Note that nothing of this payment goes to the creators of WYSIWYG Web Builder, we do not make any profit from this functionality!
We just want you to be able to use it in the software.
Are there any limitations for a free account?
Unfortunately, a free OpenAI account has limitations. For example, there is limited number of images you can create per minute. The same applies to text.
More details are available here:
Is the ChatGPT API included in a ChatGPT Plus subscription?
No, unfortunately the ChatGPT API and ChatGPT Plus subscription are billed separately.
The API has its pricing, which can be found at https://openai.com/pricing. The ChatGPT Plus subscription covers usage on chat.openai.com only.
"You exceeded your current quota, please check your plan and billing details" error
This error indicates that you have reach a limitation of the (free) account.
For example, you may have generated too much images or text per minute. In that case, try again later or upgrade your current OpenAI plan.
Note that WYSIWYG Web Builder has no control over this. It just displays the error that was received from the OpenAI server.
Here is an overview of all OpenAI errors and their meaning:
How to improve the quality of the images?
By default, the software uses the DALL-E 2 model to generate images, but DALL-E 3 is also supported which generally gives much better results.
This can be configured in Tools -> Options -> Artificial Intelligence
DALLE·E 3 is the highest quality model and DALL·E 2 is optimized for lower cost. See also: https://openai.com/pricing
How to create better prompts to generate higher quality AI Images?
1. Be specific and detailed in your prompts for better results. Asking for "a monkey" may result in odd images, but requesting "an oil painting of a monkey in the jungle" produces better outcomes.
2. Adjust your expectations; achieving impressive results may require experimentation and refining your prompts. It often takes a few tries with different variations to get the desired outcome.
3. Avoid overly complicated prompts. Including too many characters and details can confuse the AI, leading to messy results. Keep it simple and focused for optimal performance.
There is a dedicated tutorial about AI prompts.
What about privacy?
WYSIWYG Web Builder directly communicates (via a secure connection) with OpenAI. It does not store any data from your input on WYSIWYG Web Builder servers or any other third-party servers.
Local data is cached for performance reasons. Normally, this data can be found in the folder: Documents\WYSIWYG Web Builder\system\ai\
What about safety?
Please visit the OpenAI website for more information about safety.
Can I view previous generated images?
All generated images are saved in the folder Documents\WYSIWYG Web Builder\system\ai\
What languages does WYSIWYG Web Builder AI support?
AI currently supports more than 25 different languages, including English, French, German, Italian, Spanish, Portuguese, Russian, Japanese, Chinese, Korean, Arabic, Hindi, Indonesian, Malay, Thai, Vietnamese, Turkish, Greek, Hebrew, Dutch, Polish, Swedish, Danish, Norwegian, Finnish, Czech, Hungarian, Romanian, Slovak
Are there any copyright issues content generated with WYSIWYG Web Builder AI?
AI-generated images and text can be used on your website freely.
Please visit the OpenAI website for more information about terms-of-use.
Which AI models are supported?
WYSIWYG Web Builder AI currently supports, gpt-3.5-turbo, gpt-4, dall-e-2 and dall-e-3.
These modals can be configured in Tools -> Options -> Artificial Intelligence
Note that you can only use models that your OpenAi account has access to. For example, not all accounts can access gpt-4 at the time of writing.
If new models become available, then you can enter the name of the new model in the input fields.
As a software developer, why do you depend on OpenAI instead of developing all the AI functionality yourself?
The short answer:
This is impossible.
The long answer:
Creating a custom version of ChatGPT/DALE-E or a similar system independently involves several challenges and considerations. Here's a short list of reasons why it will not be feasible for an individual developer:
- Enormous Computational Resources: Training large language models requires substantial computational power, often in the form of high-performance GPUs or TPUs. Acquiring and maintaining such resources can be prohibitively expensive for an individual.
- Massive Datasets: The training of state-of-the-art language models like GPT requires access to massive and diverse datasets. Building or obtaining such datasets independently can be challenging and may raise legal and ethical concerns.
- Expertise in NLP and Deep Learning: Developing sophisticated language models demands expertise in natural language processing (NLP) and deep learning. Achieving a comparable level of performance to existing models requires a deep understanding of advanced algorithms and techniques.
- Time and Effort: Creating a high-quality language model is a time-consuming process that often involves experimentation, tuning, and iterative development. The level of effort required can be daunting for an individual developer with limited resources.
- Infrastructure for Deployment: Building the model is just one part of the process. Implementing a robust and scalable infrastructure to deploy and serve the model to users involves additional challenges, such as managing server resources, handling user requests, and ensuring low-latency responses.
- Ethical and Safety Considerations: Large language models, especially those that can generate human-like text, raise ethical concerns related to potential misuse, bias, and misinformation. Addressing these concerns responsibly requires careful consideration and safeguards.
- Ongoing Maintenance and Updates: Language models require continuous maintenance, updates, and improvements. This ongoing commitment can be challenging for an individual developer without the support of a dedicated team.
- Community and Collaboration: Large language models are often developed collaboratively by teams of researchers and engineers. The synergy of diverse skill sets and perspectives is valuable in addressing the complex challenges involved.
OpenAI, has over 500 employees and is backed by major companies like Microsoft, which invested $1 billion in the AI development company, has achieved a level that is challenging to match.
1 post • Page 1 of 1