As I work to streamline the process of adding new books, and thier contents, to the library database one overwhleming task is the raw data-entry of hundreds, thousands, tens-of-thousands of tricks found in all those pages. Let me introduce you to a serious time-saving feature.

1. Take a photo of the Table of Contents from a book
2. Upload the photo
3. Let AI "read" the page, and format the titles and page numbers for you
4. Double-check the work
5. Save the results

Very little typing required, and I've been able to enter a hundred or more items in a matter of minutes. Take a look at the feature in action...

Snap a clear photo for best results

While the AI is surprisingly good at reading strange stuff, including multiple columns of items, you will save yourself a lot of grief by starting with a good photo.

• Even lighting, with no shadows or bright spots.
• Keep the page as flat as possible.
• Square up your camera, stright above, so the page is not warped.
• Crop out any non-content text.

Here's an example of a nice clean Table of Contents ready for scanning.

Use Your Own Intelligence to Tidy Up

The result of the AI scan will be the raw text, which might look like this...

# SECTION TWO
COIN BOX MAGIC
Introduction To Section Two - 167

NINE
The Click Pass - 169

You will need to use your own brain, and knowledge of the book, to clean up the formatting a little bit before saving. You can also use the shorthand symbols to identify items as Sections (#), Tricks (!), Techniques(=), Articles (@) etc.

# SECTION TWO: COIN BOX MAGIC
@ Introduction To Section Two - 167

# NINE
= The Click Pass - 169

Overall, the theory around here is that we make progress through progressive improvement. Being able to enter the raw contents, even if there are little mistakes, is better than not at all. It's easy for users to update and correct any errors when they are noticed... and if they are never noticed then I guess it's not important!