About

A searchable list of the coffee roasters working in the United Kingdom. The list itself lives in data/roasters.csv; this page is a reader for it, and the CSV is the source of truth.

Where the rows come from

Roasters were collected from roasterlist.co.uk, an independent UK roaster directory whose robots.txt permits crawling and which publishes a sitemap. Every row records the page it was found on in source_url, so a bad source can be re-checked in bulk.

A row belongs here if the business roasts its own coffee and does so in England, Scotland, Wales or Northern Ireland. Cafés that buy in roasted beans, importers and resellers are out. So are the Crown Dependencies — Guernsey, Jersey and the Isle of Man are not the UK, and four roasters were excluded on that basis.

Where the coordinates come from

Coordinates are derived, never scraped. A scraped coordinate is usually a map pin for a café rather than the roastery, and there is no way to tell after the fact. The postcode is looked up instead, and every row records which kind of position it holds:

0: from a full postcode
0: a town or district centre
0: not placed at all

Where the directory gives no postcode, three further sources are tried in turn: the roaster’s own website, which is the business stating its own address; the Food Standards Agency’s food hygiene register, which reaches the small home-run roasters who publish no address at all; and any place named in the directory’s own prose. Each row says in notes which of these placed it.

The register gives only a postcode district for a home-based business. That redaction is deliberate, and it is respected here: those roasters are placed at the centre of the district and never resolved closer.

What is left unplaced is not an oversight. Most of it gives a region rather than a place — the Cotswolds, the Lake District, Shropshire Hills — or no location whatsoever. Town names are ambiguous, too: there are five Newports across three countries, so a name is only accepted when the county or country agrees, and left blank rather than guessed at. Where two sources disagree, the postcode wins: one roaster filed under London turns out, on the register, to be in Glasgow.

What “active” means

Every website was requested once. 0 answered and are marked active; 0 did not and stay unconfirmed, with the HTTP status in notes. A site that fails to answer is not proof a roaster has closed, so nothing is marked closed on that evidence alone.

Acknowledgements

This is built on open data, and one of those sources asks to be credited:

Contains OS data © Crown copyright and database right 2026.
Contains Royal Mail data © Royal Mail copyright and database right 2026.
Contains National Statistics data © Crown copyright and database right 2026.

postcodes.io — postcode and place lookup, free and without a key. Built on ONS data published under the Open Government Licence v3.0.
Natural Earth — the coastline, 1:50m subunits, public domain. Simplified here to 3.3 kB, which is why the map needs no tile server.
Gibbs/uk-postcodes — postcode district centroids, CC0, itself derived from ONS data. These ride along in this file, so a postcode can be found with no network at all.
Food Standards Agency — the food hygiene rating register, which placed the roasters who publish no address of their own. Crown copyright, Open Government Licence.
Roasterlist — the directory the list was built from.

The code

The scrapers, the schema, the CSV and this page are all at github.com/dolegi/bean. The pipeline caches every fetch, so re-running it costs nobody a second request.