How Brave's Privacy-Protecting AI Can Now Help You Work Faster Without Sacrificing Security

QUICK ZAPS

Brave Software announced yesterday that its privacy-focused AI assistant Leo has gained new abilities to interact with popular productivity documents and files. Leo can now understand PDF documents by accessing their underlying semantic structure through accessibility metadata. It can also extract text from images of documents in Google Docs and Sheets using optical character recognition (OCR) technology. 

This allows Leo to help users working in many common apps like Google Drive. It can propose relevant terms while editing Docs, analyze tables and suggest formulas in Sheets, or highlight key points from long Slack threads to save time. Leo's document support also extends to summarizing videos on YouTube and extracting information from deep within multi-page PDF research papers and technical documents.

Brave designed these new features while maintaining the privacy protections that set Leo apart. All requests are anonymized through a reverse proxy, and conversations are not stored on Brave's servers. No personal data is collected or used for model training. Users do not need to create an account, and subscription tokens cannot be linked back to individual activity. This ensures sensitive work and school documents discussed with Leo remain private.

The document capabilities are available now to all Brave desktop users on the latest version. Users can access Leo's help directly from PDFs, Google Drive files and more with just a click. Brave plans to expand Leo's integrations to more file types like GitHub code in the future.