You can now publish a Wiki-like page for your concept map. This page contains a summary, key takeaways, and concepts from your exploration, and it is ideal for sharing with others to provide a high-level overview of your concept map. Heuristica pages are publicly available and will soon be discoverable by other platform users.
Previously, Import and Export were performed using the same modal. There are now two different modals for these operations.
The latest state-of-the-art model, Claude 3.5 Sonnet, is now available on Heuristica!
You can now review flashcards in an Anki-like fashion using spaced-interval repetition techniques. This is currently implemented inside flashcard decks but will soon work across all your flashcards.
You can now also export flashcards from inside the flashcard decks in your library to be able to import them inside apps like Anki
You could previously export image files using Heuristica. Now you could export the map as a markdown. Additionally, you could import Markdown data exported from other mindmap tools.
Previously, Autosave in Heuristica wasn't supported. I didn't want to autosave your map in case you made a mistake that you didn't want to persist. However, now that Undo and Redo are implemented, it also makes sense to support Autosaving. After the map is saved manually, your work will autosave every minute.
A new Search with GPT option is now available under the models. Using this model, you can perform an internet search for every prompt button click. It is a great way to ensure you have references for the information generated by Heuristica. The results are not always on par with what you could be getting from the LLM, though.
You can now undo and redo actions like deleting nodes, creating nodes, moving nodes, etc.
You can now change the LLM model from the sidebar. This should allow you to quickly compare and contrast the responses you get from different models.
