Behavioral layer

Smart context.
Only what's needed.

No photos. No names. Just four behavioral dimensions that let the right person recognize themselves without exposing who they are to everyone else.

☕ Context Card
Cafe Intermezzo · Apr 17 · 9:14am
Venue
Coffee shop
Independent, not chain
Time window
Morning rush
9–11am
Crowd density
Medium
Busy but not packed
Movement
Seated, solo
Near window, table for one
The four dimensions

What every post
tells you.

Every moment on Misd carries a Context Card — four behavioral signals set by the author that narrow down who the post might be about.

📍
Venue type
Where it happened
Not the specific address — the category of space. Coffee shop. Transit. Music venue. Park. Gym. Street corner. The venue type sets the physical and social context without revealing a precise location that could identify the author.
Examples
☕ Coffee / café
🚇 Transit / commute
🌿 Park / outdoor
🎵 Music venue
🕐
Time window
When it happened
Not a precise timestamp — a window. Morning rush. Late afternoon. Late night. The time window narrows context significantly without pinpointing the author to a specific moment that could be cross-referenced with surveillance or transit logs.
Examples
🌅 Early morning (6–9am)
☀️ Morning rush (9–11am)
🌙 Late night (10pm+)
🌆 Evening (6–9pm)
👥
Crowd density
How packed it was
Empty, sparse, medium, busy, packed. Crowd density changes how likely a specific person is to be noticed — and how likely they are to have noticed someone else. A "packed" venue in a city of millions is very different from "sparse" at a niche gallery.
Examples
🟢 Empty / quiet
🟡 Medium crowd
🔴 Packed / standing room
⚡ Live event density
🚶
Movement
What you were doing
Seated alone. Moving through. Standing at the bar. Waiting in line. The movement dimension is often the most specific — it tells the right person exactly where in the space and what they were doing when the moment happened. It's the behavioral fingerprint.
Examples
💺 Seated, solo
🚶 Moving through
🧍 Standing at bar
⏳ Waiting in line
How precision works

More dimensions,
fewer false matches.

Each dimension you add exponentially reduces the pool of people who could plausibly be the post's subject — and increases the signal quality of a "This might be me" tap.

// DIM_01
Venue only
If a post says "coffee shop" with nothing else, thousands of people could claim it. The signal is vague. The noise is high.
~8%
signal precision
// DIM_02
Venue + time
Adding a time window cuts the candidate pool by 3–4×. Now the post is about people at that café specifically during morning rush — a much smaller group.
~31%
signal precision
// DIM_04
All four + anchor
All four dimensions plus a Memory Anchor reduces false positives to near zero. The only people tapping are people who were genuinely there, doing that specific thing.
~94%
signal precision
Fifth dimension

The Memory Anchor —
only they would know.

An optional verification question added by the author. Something so specific that only the actual person described could answer it — and wrong answers still cost a daily signal.

Context Cards tell the story of where and when. The Memory Anchor asks a question about the specific content of that moment — a detail only the person who was there would know.
It's not a riddle and it's not a quiz. It's a filter. You either know the answer immediately because you were there, or you don't — and if you don't, you're not the right person.
Wrong answers are silent. The author only sees that a signal was sent and didn't clear the anchor. They never learn who answered wrong or what they wrote.
Wrong answers still cost. A signal is consumed regardless of whether the anchor was cleared. This prevents people from guessing across many posts hoping something lands.
Partial matches count. "kafka" clears the same as "Kafka on the Shore" — the system checks for meaning and intent, not spelling or exact phrasing.
🔑 Memory Anchor in action
"You were reading a worn copy of Kafka at Cafe Intermezzo. I spilled my latte trying to say something. You looked up for exactly one second and I completely lost the sentence."
"What book was I reading when the latte incident happened?"
Kafka on the Shore
✓ Memory confirmed — match fires if they also signal
What makes a good anchor
✓ Specific to the exact moment — not a general fact about the place
✓ Observable — something they would have seen, not guessed
✓ Narrow enough to exclude bystanders nearby
✗ Not something Googleable or publicly knowable
✗ Not so obscure that even the right person couldn't answer it
Behavioral patterns

When context repeats,
it becomes a pattern.

Regulars at the same coffee shop leave behavioral footprints. Context Cards, aggregated over time, surface patterns that make recognition even more precise.

📍 Cafe Intermezzo, ATL
Recent context patterns
Tue–Thu, 9–10am, seated window
Solo, reading, medium crowd
Leaves before 10:30am
3 recent signals here
🚇 MARTA Red Line
Recent context patterns
Weekday, 8–9am, standing
Headphones in, facing doors
Exits at Midtown station
2 recent signals here
🌿 Piedmont Park
Recent context patterns
Weekend mornings, moving
Solo run, east loop path
Sparse crowd, early light
1 recent signal here
Context patterns are only visible on venue-level Place pages — not on individual profiles. No one can build a map of your movements from this data. The pattern shows that someone is a regular at a place, not who they are or where they go next.

Context is the
whole product.

Without it, a post is just words. With it, it's a precise signal to exactly one person — the one who was there.

Browse the feed → Privacy Design