People use open tabs as a temporary memory system because bookmarks and reading lists do not preserve intent, urgency, or visual context.
Partner pitch / 2026 product thesis
Trace
A personal web memory system that turns open-tab anxiety into retrievable, visual, intent-rich memory.
我們不是在做更好的書籤。我們在做一個讓人敢關分頁、又真的找得回來的「注意力回收系統」。
Thesis
Browsers save pages. Trace saves why they mattered.
300
open tabs are not clutter; they are unprocessed intent.
2s
capture must be faster than deciding where it belongs.
1st
wedge: iPhone Safari users drowning in valuable tabs.
01 / Problem
People keep tabs open because current tools erase context.
A tab is often not a page. It is a weak signal of desire, taste, comparison, curiosity, unfinished reading, or a future purchase decision.
Want to buyProduct pages stay open because the decision is unresolved.
Commerce
Want to rememberStores, makers, restaurants, references disappear inside history.
Memory
Want to compareTabs become a fragile comparison board with no structure.
Decision
Want to keep tasteStyle references need visual recall, not alphabetical folders.
Taste
02 / Emotional job
「我不知道現在要怎麼整理它，但我知道我不想失去它。」
03 / Why now
The timing is finally right.
Mobile browsing created the tab debt. AI enrichment can now make saved pages useful later. At the same time, people are tired of generic AI surfaces and want tools that feel personal, visual, and trustworthy.
Metadata, screenshots, summaries, visual clustering, and semantic search can turn raw links into useful memory objects.
Tactile, human, non-generic
2026 visual language is moving away from synthetic perfection toward warmth, texture, editorial hierarchy, and motion that serves storytelling.
Personal AI needs memory
AI assistants are only useful if they can access the user’s own accumulated references, taste, and half-made decisions.
User-mediated import
Trace should not secretly read Safari. The user hands pages over; Trace explains, preserves, and organizes them.
04 / Product
Trace has one core loop: Capture → Understand → Return.
The product starts as a rescue layer, but compounds into a personal web memory graph.
Capture
Safari Share Extension and batch paste rescue. Save without asking users to classify everything first.
Understand
Extract title, image, domain, summary, price clues, content type, and the user’s optional reason for saving.
Return
Bring things back through visual boards, comparison views, semantic search, and “ask my web memory”.
05 / Experience demo
The first product moment: “I can finally close these tabs.”
Click through the three entry points. The deck should sell the behavior change, not just the UI.
Capture moment
The save action must be emotionally lighter than keeping the tab open.
Trace asks for one optional reason: 想買、想比較、喜歡這個風格、有用資訊、只是先收著. Everything else can happen later.
06 / Wedge & moat
Start narrow: high-intent mobile Safari tab rescuers.
The wedge is not “everyone who saves links.” It is people whose tabs contain aesthetic references, shopping decisions, and useful research they actually want to revisit.
- Style-conscious shoppers and collectors
- People who keep boutiques, products, references, restaurants, and articles open
- They feel both attachment and anxiety toward their tabs
- Visual recall matters, so generic bookmarks fail
- Purchase/comparison intent creates repeated return value
- The “300 tabs” pain is visceral enough to motivate import
- Intent layer: why the user saved it
- Visual layer: images, screenshots, mood clusters
- Semantic layer: summaries, entities, product/store concepts
- Not a universal browser replacement
- Not a silent Safari scraper
- Not a full second brain on day one
07 / MVP
Build the smallest version that proves tab relief and return value.
MVP success is not feature count. It is whether users save, close, and later retrieve pages with less anxiety than before.
iOS capture + batch rescue
Share Extension for single pages. Paste/import flow for large URL batches. Inbox with visual cards and intent chips.
Save flow under 2 seconds
Batch import supports messy text
User can safely close Safari tabs
Does Trace create relief?
We test whether users trust Trace enough to close tabs, then whether they can find saved things days later.
70%+ imported tabs closed within session
50%+ saved pages revisited or searched
Users add intent reason to 30%+ saves
What memory views matter?
Compare cards, boards, semantic search, and weekly review. Find the first return surface that users naturally revisit.
Top recall categories by user type
Best default classification model
What must stay manual vs AI-assisted
08 / Partner ask
What we need to decide together.
This deck is not asking for agreement that the idea is beautiful. It is asking whether this is the right wedge to validate, and who owns which risks.
Product risk
Technical risk
Go-to-market risk
Next 30 days
Product risk
Will users trust Trace enough to close valuable tabs?
The first validation is emotional, not technical. If Trace cannot create relief in the first session, no amount of memory graph elegance matters.
North Star
Trace remembers the reason a page mattered.
Every feature should reduce tab pressure, preserve intent, and make return feel natural.
Product decision filter
Does this reduce the pressure of open tabs?
Does it preserve why the user kept the page?
Does it make return more likely, not just storage bigger?
Does it stay honest about Safari/iOS boundaries?