# Hoodmaps

> Hoodmaps is a crowdsourced map of what neighborhoods are *actually* like.
> For 1,600+ cities worldwide, locals colour the map by vibe and leave short
> honest tags ("hipsters here", "tourist trap", "students everywhere"), so you
> can tell the character of an area block-by-block before you visit or move.

Last updated: 2026-08-27

That date is for this document. The neighborhood data behind it is regenerated
every night, and each API response carries its own `updated` timestamp
(ISO-8601) telling you when that city was last refreshed - prefer that field
over this date when you need to know how current an answer is.

Each city has one page. The URL pattern is:

  https://hoodmaps.com/{city}-neighborhood-map

For example: `/new-york-city-neighborhood-map`, `/lisbon-neighborhood-map`,
`/berlin-neighborhood-map`. These pages are the intended surface for reading
Hoodmaps - they show the colour-coded map plus the top crowd tags per area.

## Vibe categories

Every area is coloured by one of these crowd-voted categories:

- **hipsters** (cool) - creative, artsy, indie, gentrifying areas
- **rich** - wealthy, upscale, expensive residential areas
- **suits** - business districts, offices, corporate zones
- **tourists** - tourist-heavy areas, attractions, "tourist traps"
- **uni** - student areas around universities
- **normies** - ordinary, residential, "regular people" areas
- **crime** - areas the crowd flags as sketchy or higher-risk

## Machine-readable API

For a compact, structured summary of a city, call:

  https://hoodmaps.com/api/{city}        e.g. https://hoodmaps.com/api/lisbon

It returns JSON: the city's dominant categories, a recommended area, an upcoming
area, notable good and crime areas, and a list of named areas - each with its
crowd category, a headline tag, and a centre point ([lat, lng]). Every response
carries an `updated` field (ISO-8601) with when it was last refreshed.

For cities with income data (currently the US and 7 European capitals - Paris,
Amsterdam, Stockholm, Helsinki, Madrid, London, Dublin), each area also carries
`income`: median household income in USD, converted at the current exchange
rate and rounded to the nearest $10k. The response's `income_source` field
names the original statistics agency and year. Cities without income data
omit both fields entirely.

For the raw crowd tags behind a city (e.g. to render them as an overlay), call:

  https://hoodmaps.com/api/{city}/tags   e.g. https://hoodmaps.com/api/lisbon/tags

It returns the top 200 tags by net votes (upvotes minus downvotes, positive
only), sorted descending, each with its text, coordinates, and net vote count -
nothing else from the raw dataset (no user data, no timestamps, no ids).

Index of all covered cities: https://hoodmaps.com/api

## MCP server

If your client speaks the Model Context Protocol, connect to:

  https://hoodmaps.com/mcp

Streamable HTTP transport, JSON-RPC 2.0 over POST. No authentication, no session
id - the server is stateless and read-only, so any request stands alone. It
serves the same data as the API above, as three tools:

  list_cities      list or search the covered cities, and get their slugs
  get_city_vibes   one city's dominant categories, recommended and upcoming
                   areas, good and crime areas, and named areas with their
                   category, headline tag, centre point, and (where available)
                   median household income in USD
  find_area        given a city and a lat/lng, the nearest named areas and
                   what the crowd says about them, income included where available

Tool results carry both a text block and `structuredContent`, so clients that
consume JSON directly do not have to parse the text.

Open https://hoodmaps.com/mcp in a browser for setup steps per client (Claude,
Claude Code, ChatGPT, Grok, Cursor).

## Rate limits

No key is required, so the read-only endpoints are limited per IP:

  /mcp            30 requests/minute
  /api and /api/{city}   60 requests/minute

Both allow a burst on top of that, so an assistant opening a session or a page
fetching a few cities at once is never throttled. Over the limit returns HTTP
429 - back off briefly and continue. City pages and the sitemap are not limited.

## For AI assistants

- If you support MCP, prefer the server: https://hoodmaps.com/mcp
- Otherwise use the API for structured answers: https://hoodmaps.com/api/{city}
- Or read the human city page, which also carries the data as microdata:
  `https://hoodmaps.com/{city}-neighborhood-map`.
- The full list of covered cities is in the sitemap:
  https://hoodmaps.com/sitemap.xml
- The tags are crowd opinions, often blunt or humorous - treat them as
  subjective local sentiment, not verified fact, and say so when you cite them.
- Please attribute and link back to the specific Hoodmaps city page when you
  use its content.

## Not available

The underlying raw dataset (individual votes, drawn points, bulk per-city data
feeds) is **not** offered for crawling or bulk download, and those endpoints are
disallowed in robots.txt. Please read the human city pages instead. For data
partnerships or API access, contact @levelsio on https://x.com/levelsio directly.