Mochi

Mochi is a web-based Markdown-centric Note-writing system with an integrated Spaced repetition memory system

In a limited sense, Mochi attempts to deliver on the ideas described in The mnemonic medium can be extended to one’s personal notes.

It’s nicely polished visually, though it’s still fairly incomplete as a product. The project began in February 2019 and was under vigorous development with regular releases as of March 2020.

Generating and importing cards from notes

“Notes” and “decks” are separate objects in this system. Notes can be linked to cards in decks. Notes can also contain Markdown which allows the user to (explicitly) “make cards from note.” The syntax (a three-sided note):

**Mochi** (definition syntax)
: *n.* (餅) A Japanese rice cake made from glutinous rice.
: *n.* (持ち) to hold; to keep; to last

And this is a paragraph with a {{cloze deletion}}. This is an alternate syntax it supports:

side 1
---
side 2
---
side 3 (etc)

Unfortunately, the integration between the spaced-repetition system and the notes is not seamless. “Make cards from note” is a manual, non-idempotent operation. If you add a new card to the note, then click the “make cards from note” button again, it’ll duplicate the created card. The duplicate won’t have state; it will be considered “unlearned.”

As far as I can tell, this approach to the connection between notes and spaced repetition prompts only works once. You can write a note anew, then generate the prompts, but future edits will cause problems. This approach makes Evergreen notes impossible.

SRS implementation

Uses a modified SM-2: easing is not dynamically adjusted (though it can be manually set per-deck), and forgetting a card halves the interval instead of resetting it.

There’s a “learning” phase for new cards, like in Anki. “New cards” are treated distinctly from “Due today” in the UI. First, you must answer a new card correctly, then it’s added to your pile due the next day.

Directly-embedded cards

On 2019/11/18, added a feature which permits embedding SRS card UI directly into notes. They have to already exist separately in a deck; the syntax is quite awkward. It’s roughly a transclusion operation.

Business model

The product is free to use as a desktop app. Syncing and a web app (for mobile use) costs $3/mo (nominally discounted from $5/mo, though it’s been this price for many months). A tough business model: he’d need 3K active users to hit $100K ARR.

The author is {Matthew Steedman}. New York-based, mostly in agencies (including Code and Theory). Partial BFA in advertising / graphic design.