Docstrange offers two processing models to safeguard data privacy:
- local processing mode: Runs entirely on the user's local device, data is not uploaded to the cloud. OCR dependencies (e.g. Ollama) need to be installed, which can be done via the
--cpu-modemaybe--gpu-modeparameter is enabled and is suitable for handling sensitive documents. - Cloud processing model: While a network connection is required, users can obtain NanoNets' API key to control access. Both modes ensure secure data processing.
For highly sensitive data, it is recommended to use the local processing mode, which may require stronger hardware support, but can completely avoid data outflow.
This answer comes from the articleDocstrange: a tool for extracting data from documents and images and converting them to multiple formatsThe
May not be reproduced without permission:AI productivity tools " How does Docstrange safeguard user data privacy?
English 
简体中文 
日本語 
Deutsch 
Português do Brasil