Hi AW2307, thanks for flagging this. We’ll look into Markdown performance, though we don’t have a workaround or a timeline to share just yet. Sorry for the inconvenience.
I think a short-term solution could be to only render markdown for the current viewport and a buffer around it, rather than loading everything at once on opening a map.
Even better, though maybe more complex: The processing would handle all cards at once but in a background thread, without blocking general interaction with the map. Currently, it freezes the map completely.
On a macbook, the system handles it with raw processing power, here it’s not reproducible for me. So please test on an iPhone to reproduce the behavior.
One additional angle: cache the rendered result per card, keyed by content hash, and invalidate only when that card’s markdown changes. That avoids reparsing all 6000 comments every time the notebook opens. Combined with viewport rendering and background parsing, the first open may still be slow, but subsequent opens should improve. It may also be worth profiling whether parsing or layout is the main bottleneck.
Thank you all for your feedback and suggestions! We’re working on optimizing Markdown performance to reduce freezing and improve responsiveness in notebooks with large numbers of cards.