Every screen and state, Phase 1 MVP and Phase 2 comparison, including multi-component (lined garment) handling. The result is a pure honest verdict: it tells you what the fabric is and gets out of the way, with no purchase suggestions. Lo-fi and greyscale on purpose. Result screens use the card-over-Home pattern (tap to expand, two states).
Lo-fi · complete coverage · honest verdict, no suggestions
Phase 1 · MVP
Type a fiber composition, get a fast honest verdict, expand for the why. Handles single-fiber, blends, all three verdict bands, multi-component garments, and every unhappy path. The verdict states the facts and stops there, no alternatives offered. The scan path is the fast-follow tier within Phase 1.
A · Home
Core v1the input, teaches through use, no onboarding
A1
Read the fabric, not the marketing.
Type what the label says.
e.g. 80% cotton, 20% polyester
Check
Scan a label
About
Home · empty (first run) in-field example is the only teaching
A2
Read the fabric, not the marketing.
Type what the label says.
80% cotton 20% polyester
Check
Scan a label
About
Home · typing forgiving parser, any format
A3
Read the fabric, not the marketing.
asdfgh
Hmm, we couldn't find a fabric in that. Try "60% cotton, 40% polyester".
Check
Scan a label
About
Home · invalid on submit gentle prompt, re-shows the example, never blames
B · Result, the card lifecycle
Core v1single composition (a blend). Peek to expanded to why, all tap-driven
B1
See the full breakdown ⌄
Verdict
Okay, with tradeoffs
58/100 · illustrative
Mostly cotton, but the blend sheds plastic and is hard to recycle.
Check another
reserved
B1 · Peek verdict, score, and the reason. No scroll, no suggestion
→
tap expand
B2
tap to close to Home
Okay, with tradeoffs
58/100 · illustrative
Made of
Cotton
80%
Polyester
20%
Breaks down in ~20 to 200 yrs
Sheds tiny plastic fibers when washed
On skin: breathable, mostly gentle
Why this score ⌄
Check another
B2 · Expanded (why closed) verdict pinned, facts beneath, then Check another. No better-path block
→
tap "why"
B3
tap to close to Home
Okay, with tradeoffs
58/100 · illustrative
Why this score ⌃
Climate
Water
Land
End of life
On skin
Check another
B3 · Why, inline five dimensions open in place; content below reflows. No second sheet
C · Verdict variants
Core v1all three bands, shown at peek
C1
See the full breakdown ⌄
Verdict
Good to buy
85/100 · illustrative
Natural, breaks down fast, kind to skin. Nothing to flag.
Check another
reserved
C1 · Good (100% linen) green is earned; states the facts, nothing to flag
C2
See the full breakdown ⌄
Verdict
Better to skip
43/100 · illustrative
Plastic based. Sheds fibers and lasts a long time in landfill.
Check another
reserved
C2 · Skip (100% polyester) calm clay, not alarm; an honest verdict in plain words, never shames, no alternative pushed
D · Multi-component garment (lined)
Core v1shell and lining differ; one verdict anchored on the shell, lining as honest modifier
D1
See the full breakdown ⌄
Verdict
2 parts
Okay, with tradeoffs
55/100 · illustrative
A wool shell, but the polyester lining sheds plastic and is hard to recycle.
Check another
reserved
D1 · Multi-component peek one verdict + a "2 parts" marker; the lining is already in the reason
→
tap expand
D2
tap to close to Home
Okay, with tradeoffs
2 parts
55/100 · illustrative
Made of · 2 parts
ShellWoollower impact
LiningPolyesterhigh impact
Sheds tiny plastic fibers (from the lining)
Won't fully break down · the lining makes it hard to recycle
On skin: the lining sits against you. Polyester, breathable but can feel warm.
More than one material. The label doesn't give exact proportions, so the verdict is based on the shell.
Why this score ⌄
Check another
D2 · Multi-component expanded grouped SHELL/LINING rows (no proportion bars), lining-driven flags, skin note, and the honest-limit note
E · Result, unhappy paths
Core v1where trust is won, per the messy-real-world principle
E1
tap to close to Home
Okay · on the part we read
61/100 · lower confidence
confidence: medium
We don't have data on "Cupro" yet. Here's the rest.
Made of
Cotton
70%
Cupro ?
30%
Fix the spelling
Check another
E1 · Partial (unknown fiber) score the known part, name the gap, offer a fix, never dead-end
E2
We couldn't recognise a fabric here
Check the spelling, or type the fiber content straight from the label.
Fast followon-device OCR; confirm the read, then straight to the verdict
F1
Line up the fiber content label
Upload
Type
F1 · Capture framing guide; upload or type as alternatives
→
shutter
F2
Reading the label…
F2 · Reading brief, on device, nothing uploaded
→
read
F3
Here's what we read.
Fix anything that looks wrong, then get your verdict.
Recognised text
80% cotton 20% polyester
Get verdict
Retake photo
F3 · Confirm the read (the trust screen) the recognised text, fully editable in one place; multi-part labels come through as text. Only scores after you confirm
↘
if unreadable
F4
Couldn't read that clearly
Poor light or a worn label can do that. Type what it says instead, it's just as fast.
Type it in
Try again
If the camera is blocked, Threadwise asks permission, or you can type instead.
F4 · Read failed / no camera always a way forward; type-it-in is the fallback
G · About and system states
Core v1About is a route; offline is a non-event; error is distinct from Skip
G1
‹ Back
How it works
Scores are illustrative, meant to guide, not certify.
Everything stays on your device. Nothing is uploaded.
G1 · About method, honesty cue, privacy; a plain pushed page
G2
Works offline
Read the fabric, not the marketing.
80% cotton 20% polyester
Check
Scan a label
No signal needed. Nothing here ever calls a server.
G2 · Offline works normally; a quiet chip, a feature not a failure
G3
Read the fabric, not the marketing.
80% cotton 20% polyester
Something went wrong on our side. Your scan is safe, nothing was sent. Try again?
Try again
About
G3 · Error plain, recoverable, neutral; visually distinct from a Skip verdict
Phase 2 · Compare (2 to 5 garments)
Opt-in and additive. The single-item flow is untouched for anyone who never taps "Add to compare." Items collect in a session-only tray, the compare view reuses the same verdict fields as columns, and a lined garment sits as one column anchored on its shell. Marking the best of the set is the comparison's job, not an external suggestion.
H · Collecting items
Phase 2the reserved slot activates; the tray docks in the reserved bottom zone
H1
See the full breakdown ⌄
Verdict
Good to buy
85/100 · illustrative
Natural, breaks down fast, kind to skin.
Check another
+ Add to compare
H1 · Result + Add to compare same screen as MVP; the reserved slot is now filled. Add from the peek, no need to expand
→
add
H2
Read the fabric, not the marketing.
e.g. 80% cotton, 20% polyester
Check
Scan a label
Holding to compare
2linen×cotton×
Compare
About
H2 · Tray docked appears only once holding items; session-only; chips removable
→
keep adding
H3
Verdict
Okay, with tradeoffs
58/100
Check another
Compare is full (5)
Remove one to add this. Five is the max for a clear comparison.
H3 · Tray at max (5) "Add to compare" disabled with a plain reason