Live dispensary data,

built to be called, not scraped.

What teams do with this

Three worked examples on Heavy Hitters, pulled live from the API above.

curl "https://api.fuzzy.so/api/radar/brands/heavy-hitters/products/top?limit=3"

{
  "products": [
    {
      "name": "Acapulco Gold | Ultra Extract High Purity Oil - 1G Vape",
      "doors": 94,
      "priceRange": { "min": "31.80", "median": "60", "max": "86.00" },
      "thc": "88.9",
      "category": "vape",
      "deepLink": "https://dutchie.com/dispensary/strain-stars/..."
    },
    { "name": "Pineapple Express | Ultra Extract...", "doors": 92 },
    { "name": "Blueberry Blitz | Ultra Pure Gummies...", "doors": 87 }
  ],
  "total": 3902
}

import { WhereToBuy } from "@fuzzylabs/where-to-buy";

// Replaces a hand-maintained "find a store" list with the current one —
// deep links, stock state, and product picker included.
function StoreLocator() {
  return <WhereToBuy brand="heavy-hitters" site="yourbrand.com" />;
}

fuzzy_brand_whitespace({ brand: "heavy-hitters", competitor: "stiiizy" })

{
  "whitespaceDoors": 1164,
  "targetDoors": 608,
  "competitorDoors": 1614,
  "byState": [
    { "regionCode": "CA", "doors": 390 },
    { "regionCode": "MI", "doors": 320 },
    { "regionCode": "NV", "doors": 147 }
  ],
  "examples": [
    { "name": "BudBees", "city": "Mission Hills", "regionCode": "CA" }
  ]
}

React widget

@fuzzylabs/where-to-buy is an open-source kit on npm that drops a store locator onto a brand's own site, reading the same data as the API below.

import { WhereToBuy } from "@fuzzylabs/where-to-buy";

function Page() {
  return <WhereToBuy brand="heavy-hitters" state="CA" />;
}

The kit's useTopProducts hook wraps the top-products endpoint below, and <WhereToBuy> uses it automatically for its product picker. It also injects the schema.org JSON-LD below on its own; opt out with includeJsonLd=false.

MCP

The same Radar data is available as an MCP server, for agent platforms and Claude Code sessions that want name-to-entity lookup, verified links, and bounded-token stats instead of raw REST calls. Full tool list, what each one answers, and how to connect a client live on the MCP page. The endpoint is https://api.fuzzy.so/mcp, no credential and no account.

API

Where-to-buy

GET /api/radar/brands/{slug}/where-to-buy returns the stores currently carrying a brand, or one specific product with ?product=. No API key. CORS-open, so you can call it straight from your own site's browser code.

const res = await fetch(
  "https://api.fuzzy.so/api/radar/brands/heavy-hitters/where-to-buy?state=CA"
);
const { stores, total } = await res.json();

Filters: state, city (raw retailer city, not yet normalized), page/limit (default 25, max 100). Pass ?site=yoursite.com or an X-Fuzzy-Site header so we know who's using it.

Top products

GET /api/radar/brands/{slug}/products/top ranks a brand's products by door count, with price range, THC, and a deep link per product. Good for a “shop by product” picker, feeding a product_cluster_id into where-to-buy above.

curl "https://api.fuzzy.so/api/radar/brands/heavy-hitters/products/top?state=CA&limit=10"

Schema.org JSON-LD

GET /api/radar/brands/{slug}/schema.jsonld?product= returns a ready-made schema.org Product + AggregateOffer, with up to 100 per-store Offer entries (price, availability, seller, deep link). offerCount always carries the true number of stores, even when more than 100 carry it. It reads the exact same data as where-to-buy above, so the two never disagree.

<script
  type="application/ld+json"
  src="https://api.fuzzy.so/api/radar/brands/heavy-hitters/schema.jsonld?product=<clusterId>"
/>

Endpoints

The rest of the public API surface — everything under /api/radar takes no API key. Every response carrying an issue rate reports it as a 0-1 fraction, not a percentage.

curl "https://api.fuzzy.so/api/radar/brands?search=heavy&limit=2"

curl "https://api.fuzzy.so/api/radar/brands/heavy-hitters"

curl "https://api.fuzzy.so/api/radar/retailers/dutchie-fabulous-creations"

curl "https://api.fuzzy.so/api/radar/operators?limit=1"

curl "https://api.fuzzy.so/api/radar/brands/not-a-real-brand"

{
  "type": "https://fuzzy.dev/errors/not-found",
  "title": "Brand not found",
  "status": 404,
  "instance": "http://api.fuzzy.so/api/radar/brands/not-a-real-brand"
}

API keys

Everything on this page works with no key today. That's ending: a key will be required for direct access to /api/radar/* soon — no date promised, since the flip is manual — while app.fuzzy.so stays free to browse either way. Get ahead of it now: create an account and you land on Settings → API Keys, where the key is made. Already have an account? Sign in at app.fuzzy.so instead. Then send the key on every request:

curl "https://api.fuzzy.so/api/radar/brands/heavy-hitters" \
  -H "X-API-Key: fz_live_..."

• Scopes. A new key defaults to read:radar and read:movement. read:radar covers everything documented on this page; read:movement additionally unlocks door-level movement data on brand pages. Uncheck what you don't need when you create the key.
• Expiry. Optional. Set an expiration date on a key when you create it, or leave it open-ended.
• Per-key rate limits. Once the key requirement is live, each key gets its own ceiling — a sane default, or a tighter one you set yourself at creation. Not enforced yet.
• Last used. A key's row in the dashboard shows when it last authenticated a request, so a stale key is obvious before you revoke it.

Rate limits

No hard rate limit on these endpoints, no API key required. Where-to-buy, top products and schema.jsonld carry a Cache-Control header (60 seconds on where-to-buy, 5 minutes on the other two); a CDN or your own fetch layer respecting it keeps you well clear of any limit there. The brand/retailer/operator search and detail endpoints below don't set one today, so cache those yourself if you're calling them in a loop. Once a key is required (see API keys above), a keyed request also gets its own per-key ceiling on top of caching.