There is one rule on the write page of Yewmark that nothing is allowed to break. It is older than the codebase, older than the brand name. It can be stated in eight words:
No picker, modal, or required choice between the user and the cursor.
That’s it. Open the page; the textarea is there; the cursor is in it; you type. Anything else is in the way.
Most journaling apps don’t have this rule. They have, instead, a stack of small interruptions: pick a mood before you start, choose a prompt from a list, select a tag, name the entry, click through a tutorial the first time. Each one looks innocuous. Each one is a hesitation. And hesitation in writing is the place ideas go to die — the page that asked you what your mood was first is the page you stopped writing on, because by the time you’d answered, the thought had moved on.
The rule killed a lot of features. Some of them I still have a soft spot for:
- A mood-picker requirement before the entry could save. The idea was you couldn’t claim to have written a real entry without knowing how you felt. The problem was that some entries are exactly about not knowing how you feel; making you decide first is an answer the entry hasn’t earned yet.
- A “today’s prompt” modal that surfaces on opening the journal. The idea was helpful nudging; the rule said this is a pre-write choice, so out.
- A required title. Out.
- A personality picker that asked which AI voice you wanted before writing the entry. The picker now sits in Settings; you set it once.
- A tags/categories field. Tags are useful in retrospect; they are deeply unhelpful in the act of writing. Out.
What stayed because of the rule: mood and energy are chips you tap after you save, on a card that appears alongside your entry. The AI is a button you press when you want it. Voice journaling is a button on a saved entry, or a recording surface when you explicitly ask for one. The Continue-in-chat button shows up on the digest of an entry you’ve already written, not as an option in the editor.
The pattern: every interaction the product needs from you, we delayed until after the writing is done. The cost of asking a question after the fact is asymmetric. If you’re sitting with a saved entry and a chip shows up asking “how were you feeling when you wrote this?” — you can answer honestly, or skip it, and either way the writing is already on the page. If the same chip had shown up before, the question would have shaped what you wrote next.
The other thing the rule does is make the editor look almost rudely simple. A textarea is, I think, the only writing affordance the product needs to give you for the first 95% of journaling sessions. The temptation to add anything to it has to be resisted with both hands.
This isn’t a new idea. Better writers than I am have said it for longer. But it is the kind of idea that is easy to repeat in interviews and impossible to actually hold to in a product roadmap. The roadmap pressure is one-way: more interactions, more telemetry, more features that look good in a screenshot. The rule has to be written down or it gets quietly broken.
We wrote it down.