Introducing semantic image recognition

Over the past couple of weeks, we’ve been heads down on some major updates that aren’t quite ready to ship yet. While those bigger changes are in the pipeline, we’re excited to share several smaller improvements that just went live.

Anara now analyzes images in your documents with enhanced understanding of charts and tables. Our system now processes visual data more effectively, helping you extract insights faster. Simply ask a question about any image in your library to try it out.

We also added the ability to @ mention items from your library in chat messages. Use this feature to quickly reference and interact with information from multiple files across your library.

Improvements

• Shipped a performance improvement to our PDF viewer to fix slowdown issues when viewing documents with many images.
• Improved response length when "no limit" setting is selected, addressing your feedback about short answers in the chat.
• Added six new citation formats including IEEE and Vancouver, bringing the total available citation styles to 10.
• Added support for longer videos (up to 4 hours) when they already have transcripts, eliminating the need for automatic transcription.

Fixes

• Fixed an issue where the screen would reset when reordering or deleting PDF files in the sidebar.
• Fixed a bug that prevented users from submitting questions when selecting from suggested prompts.
• Resolved an issue that caused chat messages to be sent unintentionally when clicking on notes.
• Fixed drag and drop import for .txt files that wasn't working properly.
• Improved file organization and folder navigation based on your feedback about inaccurate drag and drop interactions.
• Fixed a keyboard shortcut conflict where Ctrl+K was opening both the address bar and search menu in some browsers.