Release Notes: Agentic Study Materials, Source-Based Trees, and Canvas Artifacts

DocTree's latest release makes it easier to turn real learning material into a structured workspace. You can now start from uploaded files or YouTube videos, keep more study artifacts directly on the canvas, create cleaner diagrams, and read generated articles with better support for code, formulas, and long-form streaming content.

This release also marks a shift in how DocTree generates study material. Instead of treating each learning tool as a separate button, DocTree is moving toward a more agentic workflow: you can chat naturally, describe what you need, and the right tools can be called to create useful study material around your goal.

Highlights

• Create learning trees from PDFs, DOCX files, text files, and YouTube links.
• Chat naturally and let DocTree route requests toward the right study material.
• Keep articles, summaries, resources, quizzes, flashcards, podcasts, notes, diagrams, and curricula as first-class canvas items.
• Generate and edit Mermaid diagrams inside the learning workspace.
• Read generated content with cleaner code blocks, smoother streaming, and math formula support.
• Quiz attempts are saved more reliably, so progress is easier to revisit.

What is new

More agentic study material generation

DocTree is moving toward a chat-first learning workflow. You can describe what you want to understand, practice, summarize, diagram, or organize, and DocTree can choose the right study tool for the job.

That means the workspace is becoming less about manually picking features and more about asking for help in plain language. When a quiz, flashcards, a diagram, a summary, resources, or a curriculum would help, DocTree can route the request toward the right kind of learning material.

This is an important direction for the product: AI should not just generate more content. It should help decide which study format gives you the clearest next step.

Learning trees from your own sources

DocTree can now use uploaded documents and YouTube videos as starting material for a learning tree. Instead of only typing a topic, you can attach a PDF, DOCX, or text file, or paste a YouTube link and let DocTree build around the material you already care about.

Uploaded sources are stored with the document, split into searchable chunks, and used as context for later learning work. YouTube imports can pull transcript content and attach the source video as a resource, so the resulting tree stays connected to the material that shaped it.

This makes DocTree more useful for coursework, research, technical reading, lectures, and any study session where the best starting point is not a broad prompt but a specific source.

A richer canvas for study artifacts

The canvas is becoming a fuller learning workspace. Articles, summaries, resources, quizzes, flashcards, podcasts, notes, diagrams, and curricula can now live as movable canvas items with saved positions.

That means generated study material feels less like a set of disconnected side panels and more like a workspace you can arrange around the topic. You can keep a summary near the article it explains, place a quiz beside the subtopic it tests, or use a diagram to make a complex idea easier to revisit.

Curriculum items also received deeper canvas support, including the ability to generate content for individual curriculum nodes and save those updates back into the workspace.

Better diagrams for complex ideas

DocTree now uses Mermaid for diagram generation and rendering. Diagrams can be created from chat requests, saved to the canvas, positioned with other learning items, and edited as part of the same workspace.

For learners, this is especially useful when a concept needs structure: system flows, sequences, relationships, timelines, and other visual explanations can sit next to the written material they clarify.

Cleaner articles, code, and formulas

Generated articles now render with stronger Markdown support. Code blocks are easier to read, streamed content appears more smoothly, and math and science formulas can be shown with LaTeX-style notation.

This matters for technical and scientific learning. A generated article about programming, physics, statistics, or engineering should not lose clarity because code or equations are hard to scan. The reading experience is now better suited to those subjects.

Why it matters

This release strengthens DocTree's core promise: turning learning material into a clearer, more navigable knowledge workspace.

For users, the product now starts closer to the way people actually study. You can bring in a file, a lecture, or a video, then expand it into articles, quizzes, summaries, diagrams, and curricula without losing the source context.

For the product, this is an important step toward a more durable learning system. Source-aware generation, persistent canvas artifacts, agentic tool routing, diagram support, and reliable quiz progress make DocTree feel less like a single AI output and more like an ongoing place to build understanding.

Smoother experience

Several refinements make the app steadier during daily use:

• File uploads now have clearer size handling for free and Pro users.
• YouTube handling includes better link detection, metadata previews, transcript processing, and rate-limit coverage.
• Quiz attempts now update cleanly per user and quiz, making saved results more dependable.
• Dashboard creation flows now support files, YouTube context, and faster topic creation from one place.
• Canvas positioning logic is more consistent for generated items.
• Theme colors and shared UI primitives have been tightened for better visual consistency.
• PDF parsing and document processing have additional test coverage.
• Middleware, analytics, and noisy logging were cleaned up for a more stable app foundation.

Together, these changes move DocTree toward a more agentic learning workspace: one where you can bring in source material, chat about what you need, and let the right study tools shape the next step. The next releases will focus on stability, bug fixes, and making these workflows feel more dependable during everyday learning sessions.