Mind Maps are a new way to learn from your saved content. Open any resource and pick Create Mind Map from the chat sidebar to turn it into a hierarchical tree of topics and subtopics. You can also start from a blank mind map and grow it with AI. Right-click any node and pick Expand with AI to add child topics, or run Expand on the root for a deeper exploration of the whole subject.
You can now sync your Heuristica library to a local folder on your computer. This keeps a local copy of your study notes, prompts, and other content as markdown files that stay in sync with your library. With your content available locally, you can process it using local AI agents and tools for even more powerful workflows. You can also import and export your content as zip files for backup or transfer.
The content viewer modals for PDFs, Wikipedia articles, and YouTube videos now have a unified and improved action toolbar. You can summarize, extract key takeaways, generate concepts, flashcards, or quizzes directly from the content you are viewing — or highlight a section and run any of these actions on just the selected text.
Introduced a videos page where you can check out how learning from videos works on Heuristica. Browse video summaries, read transcripts, and check out quizzes and flashcards generated from videos.
Heuristica now generates chapters for Videos and Podcasts so you can navigate the content much faster.
You can now generate concept maps from videos, podcasts, and PDFs. Open any of these resources and create a visual concept map from its content with one click.
The AI Text Editor now supports streaming responses. Instead of waiting for the entire response to be generated, you can see the AI's output appear in real time as it writes, making the editing experience feel much more responsive.
Updated the available AI models across the board. GPT-4o and GPT-4o mini have been replaced with GPT-5.4 and GPT-5.4 mini. Claude 3.5 Haiku is now Claude 4.5 Haiku, and Claude 4 Sonnet has been upgraded to Claude Sonnet 4.6. Qwen 2.5 72B has been replaced with the much more capable Qwen 3 235B. DeepSeek-V3 and DeepSeek-R1 have been consolidated into DeepSeek-V3.1. Additionally, Gemini 1.5 Pro, o4-mini, and Kimi K2 have been removed.
You can now customize the Study Note colors.
Streaming Study Materials is now also supported inside the Concept Maps.