OpenAI Generator - Home
Introduction

     This site uses the OpenAI API to generate text and image responses to text queries. You will need your own OpenAI API Key to use the services on this site. To get an API key go the OpenAi Platform website here. OpenAI gives every new account a certain amount of free credit to experiment and to make sure it's services and features are a good fit for you or your organization before setting up a paid account. This site will undergo changes and more features will be added over time. My current TODO list is at the bottom of this page, and the Changelog is at the very bottom.

     This site is not affiliated, associated, authorized, endorsed by, or in any way officially connected with the OpenAI Platform, or any of its subsidiaries or its affiliates. The official OpenAI website can be found at https://openai.com. The names OpenAI as well as related names, marks, emblems and images are registered trademarks of their respective owners.


News

2-10-2023 - Site is now live!
After a soft opening period checking for errors this site is now live at openaigenerator.net. I wanted to make sure deletion confirmation and account deletion was in place before launching on a real domain. Now that the basic site functionality has been completed and tested, I want to focus on new features like:

  • Image Variations
  • Image Masking
  • Model Training and Management

The goal is to have full API coverage and to keep track of changes as OpenAI makes them. I also want to offer a public generator that uses my API key for training and experimental purposes in the future. That's all for now, thanks for visiting and stay tuned.


Examples

Image Query: "a cat playing in a field of grass with dogs"
Image Size: 256px
Number of Images: 3

Image Response:
genimage-20230204160837 genimage-20230204160838 genimage-20230204160839


Text Query:
"describe your humanoid form"

Max Tokens: 100
Actual Tokens: 78
Temperature: 0

Text Response:
"My humanoid form is that of a tall, slender human with pale skin, long black hair, and bright blue eyes. I have a slim, athletic build and stand at 6 feet tall. My features are delicate and my movements are graceful. I have a strong sense of style and often dress in a fashionable, yet comfortable manner."

Image Response:
genimage-20230209034747-ydr9i72l

FAQ

Q: How much does it cost to use this site?
A: The services on this site are provided free of charge but I gladly appreciate donations!

Q: How much will OpenAI charge me?
A: See here for OpenAI's current pricing information.

Q: Do you store my files/data?
A: It's completely up to you! There is a setting called "Save Generations", where the site will save all generated responses to a private directory only you know about. If you do not want the site to store any history, simply turn this option off in the settings page.
Furthermore, this site does not ask/store any personal information such as name, email, etc. Only your OpenAI API Key is needed to use the services on this site.

Q: Do you keep my OpenAI API key?
A: No. Your API Key is only saved in a session variable while you are logged in. The API Key gets sent to OpenAI's servers on every request for response generation purposes only. If you choose to save a history of generated responses (see above), the site uses a hash of your API key to identify your files and keep them private from others. You can delete all of your files at anytime.

Q: Can I have more than one OpenAI API Key?
A: Yes, you can create and delete API keys as much as you want. You can generate a key, use it once, then delete it. Please note if you choose to save generations on this site, the only way to access them is by using the same API key used to generate them. I have no way of recovering lost API key generations and generations get pruned after a long period of inactivity.

Q: Can I store unlimited files here?
A: No, our space is finite and as such every user has a quota. If your quota gets full, simply erase some of your old generations in order to reduce your quota usage. Our quota dynamically changes depending on overall usage but is usually around 128MiB a user. You can see your quota usage on the Settings page and on the Generations page.

Q: How can I use the generated responses?
A: Anyway you like! OpenAI generated responses are your creations and you can do what you please with them.
Text responses can be used for articles, poetry, books, newsletters, classroom instruction, etc. etc.
Image responses can be used as art, logos, advertisements, etc. etc.
The only limit is your imagination!
See here for OpenAI's copyright and Terms of Use.

Q: So why this site?
A: When I first started experimenting with OpenAI, I wanted an automated way of saving my queries and responses at the lowest price possible. For image generation, the API lets you choose the number of images generated. Also, OpenAI's website version of Dall-E places a watermark on all generated images which may not be wanted.

Q: How do I check my OpenAI API usage and billing information?
A: You can go here to check your stats with a daily breakdown of usage and charges.

Q: What generation models are available here?
A: For images: Dall-E 2 as this is the only available image model so far.
     For text:

  • Davinci (text-davinci-003)
  • Curie (text-curie-001)
  • Babbage (text-babbage-001)
  • Ada (text-ada-001)

Q: What size image responses are available?
A: 256px, 512px and 1024px (Note: all image responses are square.)

Q: How long are the text generated responses?
A: It's up to you. When you submit a text query you choose the response length via the "Max Tokens" setting. A token can be thought of a word or part of a word. For more on this, please see "Token" in the glossary below. Responses will never be over the Max Tokens setting.

Q: I submitted a query but did not get a response, what happened?
A: 99% of the time it is due to OpenAI censoring the query. This happens if you ask a naughty question or ask it to generate an illegal response. See OpenAI's usage policy here.
Another reason could be a simple network timeout. Try again in a few minutes.
OpenAI does not charge for these types of issues because there is no response.

Common Error Messages:
"Incorrect API key provided: KEYHERE. You can find your API key at https://platform.openai.com/account/api-keys."
"Your request was rejected as a result of our safety system. Your prompt may contain text that is not allowed by our safety system."

Q: What is the maximum length of a query and or response?
A: From the OpenAI website:
"As a rough rule of thumb, 1 token is approximately 4 characters or 0.75 words for English text. One limitation to keep in mind is that your text prompt and generated completion combined must be no more than the model's maximum context length (for most models this is 2048 tokens, or about 1500 words)".

Q: What is your Warranty / Terms of Service?
A: I am not responsible for any harm that may come to you or your computer by using the services on this site.
You and you alone are responsible for the safety of your API key. Do not share your API key with anyone in order to keep your generations private and to avoid unauthorized charges.
I am not responsible for lost generations. Please save your API Keys in a secure location and save all generations to your prefered storage medium.
If you have a problem with any of the statements above, please do not use this service.


Glossary

Query
Text that is a phrase or question that gets sent to OpenAI in order to receive a response.

Response
An answer to your query. Can either be text or an image.

Token
From OpenAI's website:
"You can think of tokens as pieces of words used for natural language processing. For English text, 1 token is approximately 4 characters or 0.75 words."
OpenAI charges you for the amount of tokens for the query and response, known as Completion Tokens. See here for OpenAI's current pricing information.

Prompt Tokens
The amount of tokens in a query.

Completion Tokens / Max Tokens
The maximum amount of tokens allowed in a response.

Total Tokens
The amount of tokens of the query and response. This is the amount of tokens you are charged by OpenAI.

Generation
A single response or collection of responses to a Query.

Temperature
The temperature controls how much randomness is in the response. If you want a simple response, keep the temperature at 0. If you want a lot of random but sensible words in the response, raise it. The temperature range is from 0 to 2, and allows fractions such as 0.5. A high temperature could be good for creative articles or even jokes.

Reason / Result Stop Reason
Reason why the job was stopped. "Stop" means end of the natural response. "Length" means it reached the max tokens length.

Quota
Amount of storage space on this site per a user. Our quota dynamically changes depending on overall usage but is usually around 128MB a user.
You can see your quota usage on the Settings page and on the Generations page.


TODO
  • Add in moderation checker for queries
  • Tokenizer test for length of query/cost estimation
  • More CSS styles/improve current one
  • Translation Framework (language definitions using lang session/browser variable)
  • Additional AI services (streaming chat, audio translation, etc.)

Changelog

v.1.4a
Completed backend administration
   Error and Message management
   Page editing and backups
   Image upload and management
Added error logging to OpenAI API responses
Code Cleanup

v.1.3.3a
Added generation time stats
Simplified Generation display results
New Plain style for Desktop and Mobile
Added Contact Us form
Code logic cleanup

v.1.3.2a
Changed Temperature to allow for fractions, instead of only whole numbers
Text completions now save stats (model, temperature, token usage) along with the query and response for reference
Updated Glossary and FAQ
Added non-affiliation disclaimer

v.1.3.1a
Moved site to OpenAIGenerator.Net
Added delete confirmation prompt for files and generations
Added option to delete entire account including settings database and all generations
Added initial backend administration support
Added error logging support

v.1.3a
Refactored code to directly use OpenAI's API with curl to keep track of evolving changes
Removed requirement of openai-php/client and it's dependencies
Limited query length display in generations area to about 100 characters
Encoded response output on generation and view to make it more readable (and safe)
Created alternative table views for mobile
Fixed mobile CSS and display code (500px and less)

v.1.2.1a
Reworked display and CSS code
Removed old style sheets
Initial work done on mobile layout

v.1.2a
Changed query input to a resizable textarea
Validates API Key with OpenAI on generation to avoid errors while still allowing users to login and retrieve their files with invalid keys
Added PHP Curl library requirement
Changed image responses from URL to base64. (site will not save generated images if not requested to)
Added random quotes about AI - Submit yours to contact [at] openaigenerator.net
Set default CSS style to Clear
Changed an example generation

v.1.1a
Changed notices to use a session variable instead of a function variable in do_output
Added notice code to CSS
Changed logic for query not entered to display a notice and send back to current page
Various typos fixed

v.1a
User account functions (login/logout/settings)
Basic Text/Image Generations
Generation (files) management page (view/delete)
User quota enforcement
.htaccess protection for private directories
Basic CSS and CSS Switcher menu added
Intro, FAQ, Glossary written