One map widget, embedded anywhere

Every listing page tries to answer the same question and almost none of them do. What is it like to live here?

What it is
Web application
Built with
Flask · MapLibre GL · PostgreSQL · Stripe · Google Cloud Run
Live at
Pathway

The short version

  • Turns any address into a curated, interactive neighborhood map.
  • Embeds into a listing page with one line of code, on any platform.
  • Live travel times, Street View and per-pin analytics for the agent who owns the map.
  • Flask, MapLibre GL, PostgreSQL and Stripe. Built as a product with accounts and billing.
The Pathway web app by Letter Box Digital: an interactive neighborhood map widget with location pins, walking times, and an embed snippet.
Pathway, live at Pathway.

Why it exists

Every listing page tries to answer the same question and almost none of them do. What is it like to live here?

The listing handles the house. Square footage, photographs, the year the roof was done. Then it reaches the neighborhood and falls back on one of three things: a paragraph of adjectives, a screenshot of a map or a walkability score with no explanation attached.

The agent showing the house knows the answer cold. Which coffee shop is worth the walk. How long the drive to the hospital takes at 8 a.m. Where the good playground is. That knowledge is the product an agent sells, and it was never making it onto the listing page.

Pathway puts it there.

The design direction

The map is the product, so everything else gets out of its way. No branded chrome, no floating panels of controls, no interface competing with the thing the visitor came to look at.

The second rule was curation over automation. Pulling every business within a mile from an API and dropping three hundred pins on the map would have been easier. It also would have produced noise, and noise is what the visitor already had. Pathway maps are hand-picked. The agent chooses what goes on the map, because the judgment is the value. An algorithm can find the nearest coffee shop. It cannot tell you which one is worth the walk.

The look came from transit maps. Transit maps are strict about hierarchy, carry a lot of information without feeling dense and never decorate. Pins read at a glance, labels stay legible at every zoom, and nothing is on the map for looks.

How it’s built

Flask, MapLibre GL, PostgreSQL and Stripe.

MapLibre draws real vector maps, which is what makes Pathway feel like an application instead of a picture. It stays sharp at every zoom and every pin opens. Each pin carries live travel times and Street View, so a buyer gets the distance, the trip and the view from the corner.

The embed had to be effortless. A map drops into a listing page with one line of code, on any platform, with no plugin to install and nothing to configure. That shaped much of the build. The widget is self-contained, loads without touching the host page’s styles and behaves the same on a custom site or a template one.

Behind it, per-pin analytics show the agent what visitors opened, and Stripe handles subscriptions. It is built as a product, with accounts, billing and a real upgrade path.

Where it stands

Pathway is live. The embed works on the platforms we built it for, the subscription flow runs end to end through Stripe, and maps are on real listing pages now.

It is early and we are treating it that way. The agents using it are writing the roadmap, which is slower than guessing and produces a better product. If you want to see it work, build a map. It takes about two minutes.

Pathway makes the same point we make about property search generally. The commodity features are settled and there is no winning them back. What is still open is everything the portals cannot do: the specific, local knowledge that belongs to the person doing the work.