Percho · Vision Doc
This is the product vision. If you are here to work on the code, start with
ARCHITECTURE.md — what every folder is for — and then
CLAUDE.md for the working rules.
Live there before you move there.
1. Vision
Percho is a video-first, immersive platform for experiencing places to live.
We believe that where you live is one of the most consequential decisions of your life — it shapes your commute, your neighbors, your kids' schools, what your weekends look like, and ultimately who you become. Yet the way people make this decision is stuck in a previous era: scrolling listing databases, squinting at floor plans and staged photos, flying in for a rushed weekend of showings — then betting years of their life on a gut feeling.
Percho changes this. Through video, we let people experience what it actually feels like to live in different countries, regions, cities, and neighborhoods — the morning light on a street, the park at dusk, the Saturday farmers market, the commute — all without leaving their couch. As they explore, we continuously learn their preferences and help them narrow down, step by step, to the neighborhood and the home where they want to live long-term or try short-term.
In one line: from browsing listings to experiencing life; from searching to being understood.
The name comes from "perch": before a bird settles, it flies over many places, lands somewhere high to see the whole landscape, and only then chooses where to nest. Percho wants to be that vantage point — and those wings — for everyone searching for their place in the world.
2. The Problem
1. People buy a way of life; platforms sell listings.
Zillow, Redfin, and Airbnb all treat the "listing" as the atomic unit of information. But nobody is really buying four walls — they're buying the feel of a street, the character of a community, a particular shape of weekend. The granularity of today's products is mismatched with the granularity of the actual decision.
2. The information that matters most is exactly what listings leave out.
Listing photos show you the kitchen. They don't tell you what the street feels like at night, who your neighbors would be, or how far you'd live from the things you care about. Today, that information can only be gathered by physically walking the area — which is effectively impossible for cross-country and cross-border movers.
3. Search-based interfaces assume you already know what you want.
Most people don't. They have vague feelings — "somewhere quieter," "a place with real street life," "near the water." Filters and forms cannot capture these preferences; they need to be observed and inferred, not typed into a box.
4. The cost of getting it wrong is measured in years.
Relocating to the wrong city or buying into the wrong neighborhood means a correction cycle of one to five years and six figures in transaction costs. Few consumer decisions carry higher error costs — and few are served by more primitive decision tools.
3. Why Now
- Video is now the default way people consume information. TikTok trained an entire generation to understand the world through full-screen vertical video. The "immersive short video + interest-based recommendation" paradigm is fully proven — it just hasn't been seriously applied to the decision of where to live.
- AI makes "a video for every home and every neighborhood" economically viable for the first time. MLS ingestion → AI video generation → automated distribution can run as a fully automated pipeline, collapsing content cost from hundreds of dollars per human-shot video to near-zero marginal cost.
- Recommendation systems are mature. Preferences can be learned and inferred from natural behavior — swipes, dwell time, saves — with no questionnaires.
- Remote work has decoupled where you live from where you work. "Where to live" has shifted from a passive outcome to an active choice, creating entirely new decision journeys: relocation, digital nomadism, "rent first, then decide to stay."
- A structural window post-NAR settlement. With buyer-agent commission structures loosening, buyers now have a self-directed exploration phase before committing to an agent — exactly where a buyer-side platform can enter.
4. What We're Building
Core experience
An immersive discovery feed built on full-screen vertical video — but for places to live. When people open Percho, they aren't "house hunting"; they're wandering the world. As they wander, the system figures out where they belong.
We deliberately do not lock down the exact interaction form at the vision level — it will keep evolving with what we learn. The direction we're currently exploring is a swipe-card model: the feed is not a single content stream but a conversation, composed of several card types mixed together —
- Content cards (Listing / Community): carry the immersive video. Swipe right to like, left to pass, long-press to peek, tap to open Explore.
- Learning cards (Preference / Trade-off / Insight): use the same swipe gesture to ask, lightly. Binary preference cards reveal taste; trade-off cards force a choice between two competing dimensions, revealing priorities; Insight cards surface the AI's read on the user, who can agree or push back — and pushback is itself high-value signal.
- Engagement cards (Challenge): guess-the-price and market trivia that make wandering playful while quietly doing market education.
The form can change; two invariants cannot: the experience must be immersive (video is the best medium for understanding a place through a screen), and the learning must be frictionless (every preference signal comes from the same natural gesture and never interrupts the experience).
This is the beginning of the README. Read the rest at the original. Read the full README ↗
