If you have an old diary or other historic document you want to share online, you know that typing up a text copy can be very time-consuming. Free AI tools may be able to help.
I recently tried using AI to extract the text from several old documents, and I’m impressed with the results. The image below shows a handwritten page from Mr A N Braudé’s wartime diary, and the text version that the AI created.
The AI only got six letters wrong out of the whole page, so even including the time it took me to correct the mistakes it was still much faster than typing up the page manually.
I’ve recorded a short video that shows how to use AI to extract text from a variety of documents, including newspaper clippings, printed pages, and handwritten notes. I also cover some of the problems I encountered, and give suggested workarounds. Click here to watch the video on YouTube.
Or if you’d like to try it straight away on one of your documents, follow these steps:
- Scan a page from the document you want to convert to text. Make sure the page is flat and well lit, scan at a resolution of between 300 and 600 dpi, and save the result as a JPG / JPEG file.
- Use your web browser to visit the free AI tool Google Gemini. [Update: Several readers wrote to tell me Google Gemini isn't currently available in Hong Kong. They recommended Deepseek as an alternative.] (Other tools such as Claude or ChatGPT also work, but I had the best results with Gemini.)
-
Sign in to Google if required, then you should see a screen like this:
- Click the ‘+’ sign, marked (A), and upload the scanned page from step 1.
- Click ‘Ask Gemini’, marked (B), and type this instruction into the box:
extract the text from the attached image
-
Your screen should look similar to this:
- Click the ‘Submit’ button, marked (C).
- Review the text it creates and check for accuracy.
Comments
AI
Hi David,
Why don't you try it with your jury lists? It should work much better than using OCR.
re: AI
Hi, you can see my attempts to use AI on the jury lists at https://youtu.be/URsHQsPZEls?si=RtJRcWMKbsBWjteH&t=718 but unfortunately it didn't turn out well.
No Google Gemini? Try Deepseek
Several readers in Hong Kong wrote to tell me that Google Gemini isn't currently directly available in Hong Kong, but that you can access Gemini via https://poe.com/, or you can use https://www.deepseek.com/en as an alternative to Gemini.
OCR.space
Hi David,
I tried the first page of the 1901 jury list on the online OCR.space website, using its most accurate Engine3.
This is the resulting text file:
**Name in Full.** | **Occupation.** | **Abode.** |
| **Anderson, George Cobban** .................................................. Marine Surveyor, Jardine Matheson & Co., ............ East Point. **Arjanee, Framjee Hormusjee,** ........................................ Merchant, N. Mody & Co., ................................. Queen's Road Central. **Arnold, Thomas** .................................................. Secretary, HK., C. & M. Steam-boat Co., Ltd., ........... 4 Albany. **Babington, Anthony** ................................................ Assistant, Sheehan Tomes & Co., ......................... Praya Central. **Bamsey, William Sidney** ............................................ Resident Engineer, Hongkong & China Gas Co., Ltd., ........ Gas Works. **Barton, John** .......................................................... Assistant, Jardine Matheson & Co., ......................... East Point. **Bell, Herbert William** ................................................... The Peak. **Bolles, John Walker** ................................................ Acting Manager, Standard Oil Company, ............... 3 Elliott Crescent. **Champeaux, Guillaume de** ........................................ Agent, Messageries Maritimes, ......................... The Albany. **Cochrane, Thomas Park** ................................................ Queen's Road Central. **Collins, Francis George** ............................................... Gas Works. **Cooke, Robert** ........................................................ Assistant Manager, HK. & W'pao Dock Co., Ltd., ........ Richmond House, Robinson Road. **Craddock, Douglas William** ........................................ Magazine Gap. **Danby, William** .................................................. Civil Engineer, ................................. 5 Queen's Gardens. **Denison, Albert** .................................................... Civil Engineer, Denison & Ram, ....................... Magazie Gap. **Dickson, Charles Wedderburn** ........................................ East Point. **Fuchs, Friedrich Hermann Arnold** ................................ Accountant, Chartered Bank of I., A. & China, ........ 1 Queen's Gardens. **Fung Wa Chiu** ........................................................ Compradore, National Bank of China, Ltd., ......... National Bank of China. **Gaskell, William Henry** ........................................ Secretary, Punjom Mining Co., Ltd., .................. Des Vieux Road. **Goddard, Frederick Day** ........................................ Marine Surveyor, Burnie & Goddard, ............... Hongkong Hotel. **Goosmann, Johann Nicolaus.** ................................ Merchant, Jardine Matheson & Co., .................. Praya Central. **Gourdin, Alliston O'Driscoll** ........................................ Poel Street. **Graham, Walter Douglas** ........................................ Manager, Wilkinson, Heywood and Clarke, Ltd., .... Hongkong Hotel. **Grote, Max Carl Johann** ........................................ Stock-broker, Heemskerk & Grote, .................. 15 Caine Road. **Hancock, Sidney** ................................................ Broker, ........................................................... 10 Queen's Gardens. **Harling, George Wilhelm Gustav** ................................ Merchant, East Asiatic Trading Co., ................. Peak Road. **Haskell, David** .................................................. Merchant, ................................................ Bellevue, Peak Road. **Haupt, Armiu** ........................................................ Merchant, Mclchers & Co., ....................... Praya Central. **Hawkins, Villiers Alweyu Cassar** ................................ Sub-Manager, Hongkong & Shanghai Bank, ........ St. John's Place. **Hill, Robert Hunter** .................................................. Queen's Road Central. **Ho Fook,** ........................................................ Assistant Compradore, Jardine Matheson & Co., .... Caine Road. **Ho Tung,** ........................................................... Caine Road. **Hooper, Augustus Sheltoff** ........................................ The Castle.
The names seem to be a lot more accurate than the AI you used. There are a few missing occupations here and there. If this is put into a table, the missing occupations should be easy to spot and corrected as the abode column would be empty for that name.
This is the result for the handwritten manuscript. It's not as accurate as your AI one, but it was done with a smaller sized copy of the original.
****** Result for Image/Page 1 ******
Thursday, 16th August : The war is over at last!!! Our spirits rose yesterday afternoon, from a low, to tops when the working parties came in with the report that Tokyo had made a broadcast of the Imperial Rescript announcing the surrender. All the Nip guards and officers were lined up at 2 p.m., and, bowing to the loudspeaker on certain occasions, listened to the speech. In the meantime, our translators were busily engaged in translating the vocalic papers which we got in through our under cover channels. This morning, with no doubt any longer in our minds the crises came when we assembled for early morning muster. When the Camp Commandant, Nada, an ugly little brute with a faceful of teeth, arrived with his interpreter, Katayama Simon and Col. "Freddy" Field together with our interpreter Cyril Ornea, met them. Simon told Nada to let Katayama take the count because they wanted to talk to talk to him. Hada, completely off his guard allowed himself to be shepherded into the Camp office, where he was told that no sooner that the war was over and before we wanted to see Colonel Tokuyama, the "Pig" at once. Nada said that the peace stories were rumours, that he had no instructions from the Governor's Office, but that he would try! to get hold of the "Pig" before evening. He was told that that was not good enough and we would expect him by 10 a.m. This fairly shook the little ape and he furlinedly went away. Field dismissed the parade which was only marred by Dickie Kerr, forgetting himself by letting go a shrieking whistler. Of course the whole Camp was getting excitement at this our first act of defiance. Barriers between the two camps went by the board, but there was
Copilot tabulated the 1901 Jury List P.1
Copilot managed to put the text file into a table form, just have to add back the few missing occupations.
It can also