Checkpoint 1 · Roadmap Reconciliation — Ship Order

Where WordSense actually is, and what moves next

Three weeks of work (225 commits) never made it onto the roadmap. This page catches you up in plain language, then asks you to make five calls and set the ship order.

5 workstreams 1 launch gate 7 decisions Tap or dictate a note
Checkpoints Latest →
Part 1 — Plain status

The five things being worked on

One paragraph each. No jargon — anything technical gets a plain-English gloss.

1. Security (locking down who can see whose data)

All four security-hardening plans were actually built back on June 12. But nobody ever ran the final check to prove they work on the live site, and you never signed off. New problem found this week: right now, if a word gets deleted or re-processed in the background system, it can silently delete a real learner's saved notes and study progress along with it — because the database is wired to cascade the delete automatically. This is cheap to fix (flip a database setting) and expensive to ignore (a real user could lose their progress with zero warning).

Anything about this status that surprises you or needs more digging?

2. The flashcard screen (carousel)

Big improvement burst here. Rating a card used to take 4–5 seconds to register — now it's instant (about 1/50th of a second). We removed 16 leftover settings that didn't do anything. Study timers now only count time you're actually active, not idle time. Hint counts are now honest (they used to lie). The little "you're learning this" badges work. What's left: one plan that needs your go-ahead (a redesigned end-of-session summary), the hint-image experiment (see #4), and you saying "yes, this is done."

Anything about this status that surprises you or needs more digging?

3. The auto-pipeline (the machine that fills in word content automatically)

The big goal — the pipeline running by itself in the background with no babysitting — has been stuck since mid-May. But: we found your own answers to a checkpoint from 2026-07-02 sitting in the database, never read into planning. You said clearly: yes, build the "top up any field for a group of words" tool; keep it on this MacBook Pro only for now (you're traveling); use your Claude subscription, not a paid API; try free Google AI first before anything that costs money. About 10 of your answers on that same checkpoint said "too technical, explain again" — so the next write-up on this topic comes with plain language and a picture of how it flows, no exceptions.

Anything about this status that surprises you or needs more digging?

4. Picture generation for hints

The old plan (using a tool called Ollama to make images) is dead — it broke back in April and was never fixed. The new plan is already running: a different tool (Draw Things, with a model called Z-Image Turbo) is generating sample hint pictures on your Mac right now. It's stuck on two clicks you need to make on your machine to turn on a setting and download the model.

Anything about this status that surprises you or needs more digging?

5. Grammar teaching (brand new)

This is genuinely new territory. We built the plumbing that lets a "card" be something other than a vocabulary word — like a grammar rule. Nine test grammar cards are already live on your own profile (they cover one small pattern in Indonesian). Mechanically they worked, but you told us afterward "it didn't actually help us learn" — so we're already redesigning them (third attempt, happening in a separate work session right now). Separately, we mapped out the whole Indonesian grammar curriculum: 126 topics, grouped into 19 families, checked against 4 grammar textbooks. 45 of those topics (all about verb prefixes) are ready to write cards for today.

Anything about this status that surprises you or needs more digging?

6. "My Cards" page redesign

A separate work session is redesigning the page where you browse all your cards — better handling for paused/retired cards (with automatic un-pausing), and better filters. Nothing written down yet; it'll get its own page here once there's something to show you.

Anything about this status that surprises you or needs more digging?
Part 2 — Your calls

Seven decisions

Tap an answer where one's offered, or leave a short note. "AI-recommended" badges below are our best independent judgment on what's right for the product — not a guess at what you'd pick.
1
C1 — Should grammar teaching become an official, tracked project (not just a side experiment)?
Ground is already laid: the plumbing is built, a small pilot is live, and the full curriculum map exists (126 topics). This would also officially confirm something you've said out loud but that's never been written down anywhere: you want grammar taught before vocabulary, not alongside or after it. We need you to confirm that's really the intent before we build a curriculum around it.
Decided by Frank
2
C2 — Should the "don't let a word-delete wipe a learner's progress" fix, plus the never-run final check, both be required before we call Security done?
Right now "Security done" doesn't include either of these. We think it should — otherwise "done" is a paper status, not a real one.
Inferred from what you said
3
C3 — Restart the auto-pipeline project using the plan you already approved (top-up tool, MacBook-only, free-tier-first), rewritten in plain language this time?
You already said yes to the idea back on 2026-07-02 — this just confirms we should go write the actual plan now, in the plain-language format you asked for.
Decided by Frank
4
C4 — What specifically needs to happen before you'd call the flashcard screen "done"? (Or — is it already done in your eyes?)
This is an open question, not a tap — tell us your checklist, or just say "ship it."
Decided by Frank
5
C5 — Learners (and you) can already flag a bad card in the app. Those flags get saved — but right now nothing ever reads them back. Should fixing that become part of the pipeline project, or stay parked for now?
This lines up naturally with the "top up any field" tool from C3 — a flag could trigger exactly that tool. But it's your call whether it's in-scope now or later.
Inferred from what you said
6
C6 — 31 alphabet/pronunciation cards already exist in the database, but there's no screen that shows them to a learner yet. When should we build that screen?
No urgency signal from you yet either way — just needs a date or a "not now."
Inferred from what you said
7
C7 — There's an old open question about what happens when a learner taps a word variant (like "running" instead of "run") to study it. It's been sitting unscoped for 3+ work sessions. Scope it now, or take it off the list?
Some of the groundwork this question needed has since shipped, so it may be cheaper to answer now than when it was first parked.
Inferred from what you said
Part 3 — The big one

What order do we do all this in?

There's one of you and several work-lanes running at once. Here's PM's suggested order — tap "Accept" or leave a note to reorder.
1
Fix the security gap + finish the check. It's the one thing that blocks having a real user at all, and it's cheap to fix relative to what it protects.
2
Decide on the Grammar project. Design energy is already flowing there in a parallel session — deciding now keeps that work aimed at an approved target instead of drifting.
3
Restart the pipeline project. You already said go — this just gets the actual plan written, in plain language.
4
Close out the flashcard screen. It's in great shape already; this is polish, not a blocker for anything else.
Part 4 — Open mic, but specific

Your planning oversight, going forward

You said you'll be doing more planning oversight from here on. Tell us what that should look like.

What do you want to see on a regular basis — a weekly one-page roadmap summary like this one? Something shorter? And separately: what do agents keep getting wrong at the planning level (not the code level) that you'd want caught earlier?