Transcription demo (single image only)
Upload, drag-and-drop or paste an image of any historical document (including tables of data) to transcribe it to text.
Please note:
- As with any AI-assisted system, there can be errors in the output - human verification is recommended
- Handwriting transcription accuracy is heavily dependent on image quality and legibility of the original script - if you can barely read it, the AI will probably struggle as well
- This website is under active development and is offered here as a demonstration only. You may see errors, or things may stop working altogther at any time!
- Please email me with your feedback, comments, observations, suggestions, etc
How does it work?
TranscribeHistory uses a two-step approach:
- First the text on the image is recognised using a commercially available OCR service. This does a great job of picking up printed text and a so-so job with handwriting.
- Next the OCR text and the original image are provided to Gemini for transcription into 'Markdown' format (which allows the model to present information in tables and mark parts of the text as headings, bold, italic, etc).
- Finally the original image and the transcribed text is shown on the page.
The combination of OCR+Gemini seems to be notably more accurate than other transcription services (eg, HandwritingOCR, Pen2Txt, NanoNets, etc) and also offers an improvement on the standard output from LLMs like ChatGPT, Claude and Gemini alone.