A generated matching card arrived on a phone with three identical
definitions (the generation guards now drop those), but there was no way
to get past a bad card from inside a session. The feed chrome (deck
study, review, mistakes and exam) now has the red "!" control: one tap
records the question, skips to the next card and shows an Undo toast.
- lib/question-reports.ts: the local list (freezmo.reports.v1, zod on
read, capped 200) is the skip state — reported cards leave every
session type, count as "not attempted" in the results instead of
wrong, and Undo takes a report back. A deck where everything is
reported stays studiable so it can be regenerated.
- POST /api/reports appends one JSON line per report to
REPORTS_DIR/reports.jsonl (shared/freezmo/reports in production, the
dev data dir in dev). Deliberately write-only: the question text is
readable on the server (that is the point — prompt tuning) but never
over HTTP, since the app has no auth. Best effort: offline reports stay
local.
- The button: a red outlined rounded square with "!", the chrome's
control size, in both feed headers (hidden on the results slide).
- config: REPORTS_DIR + RATE_LIMIT_REPORTS (20); the deploy creates the
shared reports dir; the dev unit and AGENTS.md note the new server
data location.
- Suites: phase-1 (append + validation + no GET), phase-2 (button, toast,
server POST, local copy, skip counter, hidden next session, Undo, exam
pool). SW -> freezmo-v38.
A generated Czech literature deck reached a phone with three identical
matching definitions — the card could not be solved. The pipeline only
deduped matching terms (and ordering items), so repeated definitions and
options sailed through.
- parseDeck now rejects any card whose options, items, terms or definitions
repeat (case-insensitive), through the same drop-or-repair path as the
ordering giveaway guard: the card is dropped, or a deck left with too few
cards gets one repair round with a clear reason.
- The matching dedupe covers definitions too, and both it and the ordering
one only trim when the card still meets the schema minimum afterwards —
a shorter card would fail validation on the next read and take the whole
deck down with it (latent data-loss bug).
- The generation prompt states the uniqueness rule.
- phase-1 suite: a stub Ollama returns a deck with a repeated-definition
matching card (the reported shape), a repeated-option card and a
salvageable one; asserts the first two are dropped, the third is trimmed
to three pairs and the good cards ship (43 checks).
- README: decision 51.
Server-side only (lib/ai.ts, lib/prompts.ts) — no client bundle change, no
SW bump.
"Freezmo uses AI and can make mistakes. Double-check important info."
(plus a natural informal Czech translation), rendered by a shared
AiDisclaimer so the copy, style and translations live in one place.
Placements — only where the model's words reach the screen:
- Chat: under the composer, below the context/model chips — always (chat
is pure model output, so it also shows in the empty state).
- Make my homework: under the start CTA (desktop) and the mode cards'
helper line (phone), plus once under the generated result for all three
modes (make / check / guide).
- New deck wizard: under the summary helper (phone step 3) and the desktop
CTA, next to Freeze my deck.
- AI grading (short answers): under the graded feedback; study, review and
exam all share it.
Stored, already-generated content (deck page, study cards, explanations)
deliberately carries no note: the disclaimer belongs to the moment of
generation, not to browsing. Style: --text-subtle (documented at 5.6:1
light / 8.6:1 dark), 11.5px, centred, wraps to two lines at 360px, not
clickable, no icon.
Suites: chat (composer, phone + desktop), wizard (deck step 3 across
360/390 x en/cs x dark/light, homework start cards), phase3 (homework
trigger + result, phone + desktop), phase2 (stubbed grading feedback).
SW -> freezmo-v37. README: decision 50 + chat composer note.
Also fixes the deploy script's test instance: it now execs the server so
the cleanup kill reaps the real PID (the wrapper subshell left orphans
squatting on the test port — three from earlier sessions were found while
verifying this task), and the readiness poll checks the port serves the
NEW commit rather than just answering, so a leftover instance can never
be tested silently.
The content offset was fine — the rail's glass panel was the problem. As a
flex-1 item in the row-flex aside, the panel's automatic minimum size
(min-width: auto) is its content's min-content width. The chat list's
search input (an <input> defaults to size=20) has a ~192px intrinsic
width, which with the New chat button stretched the panel to 345px — 89px
past the 268px rail — where it painted over the chat page: the header
title, the composer placeholder ("sk about your notes...") and the
context chip ("ts").
- min-w-0 on the panel so the rail width always wins and wide content
truncates instead (the input and list rows already shrink/truncate).
- Chat suite: assert the panel stays inside the rail and the main content
starts after it, expanded and collapsed (the old checks counted
elements, never their boxes — this is what slipped through).
- README: history bullet + decision 49. SW -> freezmo-v36.
Verified on the dev instance and a production build: panel 244px inside
the 268px rail (60px in the 84px collapsed rail) at 1024-1920, content
starts after the rail, all sidebar pages unchanged, dark + light, and
the full e2e (12/12 suites, 516 checks).
- Production now runs from ~/freezmo-prod/current (release symlink) as a
standalone server on :3207; releases are exported by scripts/deploy.sh,
built and e2e-verified on a temp port, switched atomically, health-
checked via the new /api/healthz (commit + build time) and rolled back
automatically on failure. Last 5 releases kept; scripts/rollback.sh for
manual rollback. Shared env + data live in ~/freezmo-prod/shared/ and
are symlinked into every release (existing sync/push data migrated and
verified identical).
- Dev runs from ~/freezmo via freezmo-dev.service: next dev on :3217
(LAN only), its own build folder (.next-dev via NEXT_DIST_DIR), its own
data (~/freezmo/.freezmo-dev), DEV badge, dev favicon/theme colour,
"Freezmo Dev" PWA identity and never the production service worker.
- /api/healthz + the Settings Details version row (short commit + build
time). output: standalone in next.config (next start still works).
- AGENTS.md documents the workflow rules: only dev updates during a task,
temporary dirs/ports for test builds, and deploy-at-the-end as the only
production path. README gains an Environments and deploy section.
- Fix found by the rehearsal: lib/dev-mode.ts must not be a client module
(a server component importing a constant from a "use client" module
gets a truthy proxy, which baked dev metadata into production builds).
- SW -> freezmo-v34.
- Conversations in a versioned IndexedDB store (freezmo-chat v1, zod
validated with the same schema backup/sync use, capped at 100 x 300).
Settled turns persist (question always; answer when produced, partial
stops included); ?c=<id> keeps reloads on the conversation via
history.replaceState (no router round-trip mid-chat).
- History UI: desktop column, phone sheet (spring, backdrop, Escape),
auto titles from the first question, inline rename, delete with an
Undo toast, diacritic-folding search over titles and message text.
Scope and model are remembered per conversation; history is readable
offline.
- Assistant actions: Turn into flashcards (New deck with the answer as
notes, landing on the options step - deck seed gained step) and Save
to notes (appends to the single scoped deck, trimmed at 40k).
- Chats travel in the backup export (preview counts them) and in the
encrypted sync payload; still never the key.
- Motion from lib/motion.ts: message fade-up, sheet springs for the
model/context/history sheets, 150ms composer height transition, all
skipped under reduced motion.
- Tests: chat suite 63 -> 89 checks (history, reload, rename/search/
delete/undo, actions, export, two-device sync round-trip, offline,
desktop panel). Full set: 12/12 suites, 513 checks. SW -> freezmo-v33.
Also: tests/model-compare.mjs benchmark (diagnostic) for deciding the
default text model on numbers; defaults unchanged.
Entry points (no sixth tab): Ask in the Library header (desktop button,
phone icon), Ask about this deck in the deck menu, Ask a question on
study/homework results (compact result travels as context), and a Chat
row in the desktop sidebar. Phones get a full-screen route with back +
New chat; the tab bar steps aside.
- Client-side retrieval: chunk scoped decks (~800 tokens), BM25 against
the question + recent turns, top chunks under ~7k tokens, each
labelled with deck + section. Notes come from the deck's stored
sourceText (already kept for Regenerate); decks without it fall back
to generated content and the UI says so.
- Stateless server: /api/chat validates against the CHAT_MODELS
allowlist, trims the history to ~6k tokens, caps context at ~9k and
streams token/thinking/done/error NDJSON. /api/chat/extract reduces
attachments to text through the existing vision step.
- UI: user bubbles + full-width assistant text, streaming with stop,
auto-scroll that pauses with a jump-to-bottom button, safe markdown
(React nodes, no HTML injection; KaTeX dynamically imported only when
a message has math), Thinking row for reasoning, copy/regenerate/
edit-and-resend, model + context pickers (last model remembered),
composer with visualViewport keyboard handling, safe-area, 16px
textarea, Enter-to-send on desktop.
- Usage-limit failures keep the draft and show the shared message with
Settings/Ollama actions; stopping maps the reader abort to a clean
stopped state; offline disables sending with a reason.
- Settings: Keep notes text for chat toggle; deck menu: Delete notes
text. RATE_LIMIT_CHAT + CHAT_MODELS in .env.example and README.
- Tests: chat suite (63 checks, stub Ollama) wired into run-all.
Full set: 12/12 suites, 487 checks. SW -> freezmo-v32.
Also: tests/model-compare.mjs (diagnostic) benchmarks two models on the
same notes (latency, repair rounds, activity types) so the gpt-oss vs
gemma default can be decided on numbers; defaults unchanged.
- Orbs rise from below the bottom edge to above the top edge, each with
its own 36-68s pass (slower on phones), a negative delay (scene already
populated on load), own horizontal position, a gentle sway and a slight
scale drift; opacity fades in/out at the edges so the loop never pops.
- Only transform + opacity animate, each orb on its own composited layer
(will-change: transform); no filter/backdrop-filter, nothing animates
size/offsets/shadow. The glacier blobs are untouched.
- Opacity lowered ~38% in both themes: 0.55->0.34 light, 0.8->0.5 dark,
phone overrides 0.32->0.2 / 0.42->0.26.
- prefers-reduced-motion: rise removed, orbs return to the old static
placements (a paused-from-start CSS animation is not applied in Chrome
and would park them below the fold); tab hidden: a visibility listener
parks the whole scene (.ice-paused), blob drift included.
- Measured with a CDP perf trace on a 4x throttled phone profile: p95
33.4ms, 0 frames over 34ms, no long tasks - identical to the pre-change
baseline; ~0.6% extra main-thread CPU. Desktop stays compositor-bound
from the pre-existing blob drift, as before. tests/bubble-perf.mjs
keeps it repeatable.
- New bubbles suite (34 checks) in the repo harness; the axe contrast
matrix is now 16 runs (360/390/1280 x dark/light x EN/CS) and passes.
- SW cache -> freezmo-v31; README decision 46.
Full suite: 11/11 suites, 424 checks.
The browser harness lived in /tmp/opencode and an overnight reboot
wiped it. It now lives in tests/ and survives reboots, is versioned
and runs with one command.
- tests/lib.mjs: shared helpers (Chromium auto-detect, seeding,
IndexedDB fixtures, stub/extra instance spawn with process-group
cleanup, check/finish).
- tests/run-all.mjs + npm run test:e2e [filter]: runs the ten suites
in order with a summary; suites that need a stub Ollama or VAPID
keys start their own next start instance on 3208/3209, verify the
port is free and kill the group on exit. Test scratch in
tests/.tmp/ (gitignored).
- Suites rebuilt and green against the current build: regression 38,
mobile 94 (tab matrix, ... sheet, settings sheet, axe), wizard 83,
native 31 (PWA/offline/share/back/undo/haptics/draft), i18n 13,
freeze-loading 6, review 15, phase1 36 (backup/sync/usage limits),
phase2 42 (learning engine/reminders), phase3 21 (homework check).
379 checks, 10/10 suites in one full run.
- playwright-core + axe-core added as devDependencies; CHROME env or
the newest ~/.cache/ms-playwright chromium is used.
- README: new decision 45 documenting the harness; decision 44's
verification note updated.
The older suites' exact sources were lost with the wipe, so the
rebuilt versions reconstruct the same feature areas rather than
being byte-identical.
