When documenting APIs or forms, most of the information in the fields' descriptions can be derived from the field names and types. This enables tech writers to use AI to generate this content and later review the text, filling in the gaps, which makes the whole process faster.

With this automation, you can enter field names and types to the spreadsheet and automatically create API fields tables with AI generated descriptions in Markdown or HTML. This works for multiple tables at a time if you are working with complex data structures.

See below how to set up and use this tool. You can also see an actual example in which I used this tool to create a public doc.

You don't need to know how to code, but you must have an Open AI "pay as you go" account. More details below.

There are two steps to get this tool up and running. First you must Get an Open AI API key and then set up the Google Apps Script.

You must pay to use Open AI's APIs, including ChatGPT, which is a depenency for this tool. Open AI's "pay as you go" account model allows you to pay by request, while working with a range of APIs, including Dall-e, ChatGPT and GPT-4 text completion.

This tool works with multiple tables at a time, but calls the ChatGPT API once for each table. I recommend you check Open AI API pricing, which may vary if you tweak the parameters. However, with this tool's native settings, I have generated dozens of tables for under $ 0.50. You can also set a monthly spending limit.

See this article on how to create an Open AI account and get your API key.

Once you have your Open AI API key, follow these steps to set up the script:

1. Create a new doc in Google Drive. This doc will receive the generated tables from the script.
2. Create a new spreadsheet in Google Drive.
3. In your newly created spreadsheet, go to Extensions > Apps Script.
4. This will open a new Apps Script project in a new tab. Create the necessary files and copy the code according to this repository's Google Apps Script folder.
5. Go to the Configs.gs file of your Apps Script project. And fill in the variables (see this example):
   • OPEN_AI_KEY: your Open AI API key.
   • FIELDS_SHEET_ID: ID of your newly created spreadsheet. You can get this ID from the spreadsheet's URL, which is in this format: https://docs.google.com/spreadsheets/d/+{spreadsheet-id}+/edit#gid=0.
   • OUTPUT_DOC_ID: ID of your newly created doc. You can get this ID from the doc's URL, which is in this format: https://docs.google.com/document/d/+{doc-id}+/edit.
6. Save your Apps Script project.
7. Go to your spreadsheet and fill in the first four cells of the first row with the following values:

With these steps you should be ready to go.

Learn more about Google Apps Script

Consider these URLs:

• Spreadsheet: https://docs.google.com/spreadsheets/d/1hPdQBb-7yDCvsd8XAqPg2345uigLKQi0Cvj1VwGoZs/edit#gid=0
• Doc: https://docs.google.com/document/d/1W7erIcdFKqVc33mRluih11l16CS_ZQfhC-lElfv4KGiM/edit

And this Open AI API key:

• sk-6H234wdfUwiTfdbCT3Blf45ZHzsj4wrev

Your Confgs.gs file should look like this:

OPEN_AI_KEY = "sk-6H234wdfUwiTfdbCT3Blf45ZHzsj4wrev"
FIELDS_SHEET_ID = "1hPdQBb-7yDC9nzb4e8XAqPgvSThSLKQi0Cvj1VwGoZs"
OUTPUT_DOC_ID = "1W7erIcdFKqVc33mRgQAwl16CS_ZQfhC-lElXRlWKGiM"

To generate your automatic API docs tables, follow these steps:

1. Open your spreadsheet.
2. Enter your data, where each line corresponds to a field and must contain three charateristics:
   • Table name: name of the table that field belongs to. This will be the title in the end result. This is specially relevant when documenting objects within objects.
   • Field names: name of the field.
   • Type: type of the field.
3. Choose the table format, in cell D2. Currently Markdown and HTML are available.
4. Go to the 📝 Generate menu. If you can't see this menu, reload the page and try again.
5. Click Generate descriptions.
6. Go to the output doc.
7. Copy the generated tables and use them in your doc.

ATTENTION! The tables format is hard coded, but they go through the ChatGPT API before the result is available, which may introduce some inconsistencies. So you must double check your tables formatting and review all descriptions before publishing. From the tests so far it seems unlikely that table format will be an issue, but is is very likely you will need to edit a small amount of the descriptions.