Memo 8.3 — Math upgrades for notes, PDFs, and practice tests

Public changelog

Since our last changelog on April 5, we have shipped a lot across Memo: better math rendering, a new PDF viewing experience, smoother source panels, fuller practice-test flows, cleaner usage visibility, faster notebooks, and steady improvements to Learn, Copilot, Studio, and mobile.

The headline for Memo 8.3 is math. Memo now handles LaTeX much more naturally across notes, practice tests, generated content, and exports. Around that, we also did a broad reliability pass so larger notebooks feel lighter, source navigation is more predictable, and study sessions have fewer interruptions.

Highlights

• Math and LaTeX now render in notes, generated study material, practice questions, answers, explanations, and exports.

• The PDF viewer has been upgraded with a new viewing mode, better zoom controls, smoother scrolling, and more reliable citation jumps.

• Notebook and deck pages now load less heavy content upfront, which makes large source-heavy workspaces feel faster and use less memory.

• Practice tests are more dependable, especially for math-heavy questions, full-screen study, mobile scrolling, grading, and saved results.

• Source panels are cleaner when collapsed, easier to reopen, and more stable when sources are still processing.

• Usage and plan screens are simpler, with less noise and clearer upgrade messaging.

• Learn mode, talking flashcards, Track, Copilot, Studio, and mobile all received meaningful quality-of-life improvements.

Math and LaTeX support

Memo 8.3 adds a much stronger foundation for math-heavy learning. Notes now support LaTeX rendering directly in the editor, including generated notes and manual edits.

You can now:

• Read formulas in generated notes without raw LaTeX getting in the way.

• Insert LaTeX blocks from the editor slash command menu.

• Edit LaTeX blocks inline inside notes.

• Render common inline math such as inequalities, integrals, subscripts, superscripts, and Greek letters.

• Preserve LaTeX when exporting notes to Markdown or PDF.

• See math-aware rendering in practice test questions, MCQ options, and grading feedback.

This matters most for math, physics, engineering, medicine, chemistry, statistics, economics, and any workflow where formulas need to stay readable after Memo generates or exports content.

PDF viewer upgrades

PDFs are central to many Memo notebooks, so we shipped a larger PDF-viewing upgrade in this release.

New PDF Viewer with scroll and search support

What changed:

• Made the newer PDF viewer the default for a smoother reading experience.

• You can now search keywords within the PDF Viewer

• Added better zoom options, including Actual Size, Page Fit, and Page Width.

• Improved page-jump behavior when opening citations and source badges.

• Improved viewer layout so the PDF pane behaves better inside Memo's source panel.

The goal is simple: when you click a citation, Memo should open the right source, stay focused, and not fight your current reading position. You can enable the new viewer in settings now.

Faster notebooks and lower memory pressure

Memo 8.3 includes a significant memory and loading pass for larger notebooks. Instead of loading every heavy panel and full source body immediately, Memo now loads more of the deck experience only when you need it.

Large notebooks should feel less heavy, especially when they contain long PDFs, many sources, generated Studio outputs, or large Track lessons.

Updated Podcast Voice and Realistic Cantonese Podcasts

Podcast voices have been upgraded with realistic Cantonese voices with 2 amazing Cantonese co-hosts to create a “realistic radio podcast effect”. Here is a quick demo:

Cleaner source panels

The source panel is now more comfortable to use on both desktop and mobile.

We improved:

• Collapsed source and Studio panels, which now behave more like shortcut rails.

• Tooltips and shortcut behavior when panels are minimized.

• Back behavior when reopening and collapsing panels.

• Source loading states while uploads and background jobs are still processing.

• YouTube source failure messages when transcripts are unavailable.

These changes make it easier to keep your workspace focused without losing access to the sources behind your answers.

Practice tests are stronger

Practice tests received one of the largest reliability passes in this release.

New and improved:

• Math rendering in questions, options, and grading feedback.

• Full-card mobile scrolling so long questions and answers are easier to read.

• Cleaner full-screen practice test behavior.

• Fewer duplicate or confusing marks displays.

• More reliable result saving after grading.

The result is a practice-test flow that is better for long exams, math-heavy questions, and mobile review sessions.

Notes and standalone writing

Memo now has clearer entry points for creating standalone notes, not just source-based notebooks.

Standalone full screen notes

You can now create a note from:

• The dashboard header.

• The create card.

• The sidebar.

• The new-deck flow.

Standalone notes open directly into a note-focused deck view, while notebooks still guide you toward adding sources. Notes also have distinct dashboard styling so they are easier to tell apart from source notebooks.

Learn mode and flashcards

Learn mode continues to get more comfortable for day-to-day review.

Improvements include:

• Better layout on short screens and small devices.

• More room for the flashcard during smart grading.

• Better website and text source excerpts, now rendered as Markdown instead of raw syntax.

• More reliable excerpt highlighting, even when source text formatting differs.

• Toggle-view flashcard scrolling fixes.

• Restored scheduling badges in Toggle view.

• Reduced PDF thumbnail memory retention in Learn.

Small changes like these add up when you are reviewing a lot of cards in one session.

Copilot and chat reliability

Copilot is now more resilient with large notebooks and older chat histories.

We improved:

• Handling for older or malformed persisted chat messages.

• Truncation of oversized prompt context before it hits provider limits.

• Safer handling of reasoning/tool-call message parts.

• Fewer long-running retries that can lead to server timeouts.

• Better behavior when RAG context is already available.

• More stable source and citation navigation from chat.

• Collapsible long user messages in chat.

This release focuses less on flashy Copilot UI and more on making long sessions and large notebooks behave reliably.

Studio and generation reliability

Studio generation is more forgiving when real-world edge cases happen.

We improved:

• Studio tasks now handle deleted outputs mid-run more gracefully.

• Expected non-fatal outcomes are treated as skipped instead of noisy failures.

• YouTube transcript issues now show clearer source-level messages.

• Source ingestion failures are easier to understand.

• Infographic image generation shows a clearer loading spinner.

• Notes generation and source progress states are more stable.

These changes reduce cases where a background job looks broken when Memo is actually handling a recoverable state.

Usage and plan screens are simpler

We simplified how usage limits are shown across Memo. The product now uses broader usage language instead of exposing every internal unit, model cost, or action-specific calculation.

Updated areas include:

• Pricing page copy.

• Usage settings.

• Upgrade modals.

• Copilot model picker messaging.

• Paywall and limit-reached messages.

• Onboarding and profile defaults.

The goal is clearer product communication: you should know when you are close to a limit or need to upgrade without needing to understand Memo's internal accounting model.

Internationalization and public pages

We continued expanding Memo's international experience. Memo’s user interface is now available in English, Traditional Chinese, Korean, Japanese and Spanish.

Memo is increasingly used by learners and teams across different languages, so we are continuing to make the public and product experience more consistent.

Reliability fixes and polish

Other fixes and refinements shipped in Memo 8.3:

• Settings now autosave with inline save feedback.

• Settings updates are patch-based so one section is less likely to overwrite another.

• Profile updates refresh shared UI more reliably.

• Account deletion now has a clearer completed state.

• Storage cleanup after source and deck deletion is faster and more resilient.

• Auth redirects preserve the page you were trying to open after sign-in.

• Universal search opens flashcard results more reliably.

• Add-source failures are easier to trace and recover from.

• Landing page hero loading is faster.

• Podcast and speech defaults were improved, including Cantonese podcast work.

• Mind map and fullscreen behavior received fixes.

• Billing and subscription status handling received polish.

Closing note

Memo 8.3 is a broad quality release with a clear theme: make Memo better for serious study material. Math should render properly. PDFs should open where you expect. Practice tests should save and grade reliably. Large notebooks should feel lighter. Source panels should stay out of your way until you need them.

Thank you for the feedback since Memo 8.2. We read it, we ship from it, and this release is a direct result of that loop.

Happy exploring and learning,
Jason
Co-founder of Memo