Changelog | Page 4

You can now define custom instructions to customize the responses you get from the LLM model.

You can now click on the Concepts button on a YouTube video, Podcast, Wikipedia article or a Web link to extract interesting or unique concepts mentioned in them. Previously, this was a much more tedious process and required you to extract snippets manually. Now, AI can help you with this.

Interesting and Controversy buttons now return much more comprehensive information.

Added the Qwen 2.5 72B model, which fares much better at math-related problems than other open-source models.

Removed some of the old models like GPT-3.5, GPT-4, or Claude Sonnet 3 to make room for the new ones. Additionally updated the Claude Sonnet 3.5 version to the latest one.

You can now create an essay or an essay outline from selected nodes (or from all the nodes if none are selected).

Updated the Explain button with new, useful prompts, such as Explain from First Principles, Explain with a Story, Explain Step-by-Step, and Explain Through Frequently Asked Question

You can now chat with your concept maps or the selected nodes! This is useful when you want a more interactive experience with a concept map, like asking the AI to quiz you on what you've learned. As a result of this addition, I moved the bottom dock to the lefthand-side panel and removed the Concepts and Custom Prompts features from the bottom dock as they are no longer useful.

You can now run custom prompts on multiple selections. Previously, only single node selections were allowed.

Fixed a bug where Auto Save was interrupting the generation of long-form content like summaries, essays, etc.