By 
A few weeks ago, Ropsie shared how she transformed her old French notes into new digital ones. She planned to type out the scanned handouts but hadn’t gotten around to it for weeks. So, I decided to help her out.
To save time, I looked for a way to convert the scanned images into text, avoiding the need to type everything manually. First, I exported the scans from Notability as images. The app zipped the five images, so I saved them to Files, unzipped them, and then saved them to Photos. Working with images from the Photos Library is quicker and easier. With that done, I was ready to extract the text from these images.
The first app I tried was CollaNote, a free and accessible option for anyone to try. Unfortunately, it didn’t recognize much of the text. On the first image, it couldn’t detect any text, and on the other pages, it only recognized a few sentences. So, CollaNote wasn’t suitable for what I needed.
Next, I tried Scanner Pro, which managed to extract text from the entire page—a significant improvement over CollaNote. However, it still had too many errors, and since I don’t speak French, I wanted to minimize mistakes as much as possible.
I then turned to Live Text, which picked up most of the text but missed numbers 9 and 10. To evaluate its accuracy, I copied the text into Craft, a note-taking app, and noticed a few errors right away. In a last attempt to get a more accurate extraction, I tried Google Keep.
Google Keep’s text extraction was the most accurate. While organizing the extracted text can sometimes be a challenge with Google Keep, in this case, it delivered the best results, with no obvious errors. Comparing the outputs, Google Keep outperformed both Scanner Pro and Live Text, making it my go-to app for extracting text from images.
For editing the notes, I chose to use Craft, a styled text note-taking app available for Mac and iPad. They are also working on a web version. To speed things up, I switched to my MacBook. After checking for spelling errors, I rearranged the notes to resemble the original handout. Unfortunately, Craft doesn’t allow underlining, so I had to bold the text instead, which wasn’t quite the same.
After editing and styling the notes to match Ropsie’s preferences—keeping all her French sentences in pink—I had to type out three of the five pages. The structure of these pages made it difficult to extract usable text, but since they didn’t contain a lot of notes, it wasn’t too bad. Craft had a few limitations, which meant the notes didn’t look exactly like the originals in the scans.
Overall, Craft is simpler to use than other word-processing apps, especially for adding basic, quick styles. However, it doesn’t show where your pages begin or end, which can cause issues when exporting. When I exported the notes at the Regular font size, some pages became two pages long, so I had to make the text smaller.
To ensure all the notes fit on one page, I exported each page separately and then airdropped them to Ropsie’s iPad for final touches. Some notes didn’t fit in Craft due to feature limitations or page space, so I added them in Notability, where I also underlined words I couldn’t in Craft. Since Ropsie created her notes in Notability, I had to finish the process in that app.
One page had a lot of handwritten notes, and I was impressed by how well Google Keep converted the handwriting to text directly from the image. If Ropsie wants to keep the handwritten style, she might use sticky notes to cover the text. I also made the questions stand out from the answers to keep things organized.
I did my best to make the new notes look better than the scanned versions, as text is easier to interact with than images. Knowing how particular Ropsie is about her notes, I didn’t delete the original scans, just in case she prefers them. If she doesn’t like the new version, she can always revert to the scans.
This whole process took me about five hours, mainly because I recorded the entire process. The actual scanning and editing only took a few minutes per page, which was efficient overall.
