Limits and caching
What each plan includes, how many requests are allowed and how to cache responses.
The API is built and tested. Hosting at api.garanify.com is in preparation; until then it runs locally. Keys are available on request.
Plans
| Free | Pro and Agency | |
|---|---|---|
| Notice (every page, every language) | unlimited | unlimited |
| Products with a GARAN label | up to 25 (variants count with their main product) | unlimited |
| Rules (brand, category, tag) | – | yes |
| Bulk upsert (up to 1000 per request) | – | yes |
| Webhooks | yes | yes |
When the Free limit is reached, the API only rejects the next label (garan_limit_reached). Existing labels always stay visible, even when Pro expires; stored rules keep working. Mandatory displays never disappear because of us.
Prices on the pricing page.
Rate limits
| Access | Limit |
|---|---|
| Secret key | 600 per minute and key |
| Bulk upsert | 30 per minute and key (so up to 30,000 products) |
| Publishable key | 300 per minute, key and visitor |
| No key (notice, languages, fonts) | 300 per minute and visitor |
Every response carries RateLimit-Limit, RateLimit-Remaining and RateLimit-Reset (seconds). Above the limit you get 429 with Retry-After. Wait at least that long, then retry.
HTTP/1.1 429 Too Many Requests
Content-Type: application/problem+json
Retry-After: 17
RateLimit-Limit: 600
RateLimit-Remaining: 0
RateLimit-Reset: 17 Caching
The public endpoint and the renders are built for caches: browsers and CDNs may serve them fresh for 5 minutes, revalidate in the background for a day and show the last response for a week if the API is down. With If-None-Match you get 304 without a body.
Fonts under /v1/assets/fonts/ are immutable. Management endpoints with a secret key send Cache-Control: no-store.
HTTP/1.1 200 OK
Content-Type: application/json
Cache-Control: public, max-age=300, stale-while-revalidate=86400, stale-if-error=604800
ETag: "b1f0c3…"
Vary: Origin
GET /v1/public/shp_4kq8Zt2mWc9x/products/WCR870?key=gfy_pk_…
If-None-Match: "b1f0c3…"
HTTP/1.1 304 Not Modified Safe retries
PUT and DELETE are safe to retry by design. For the creating POST endpoints (…/products/bulk, /v1/shops, …/keys, …/keys/{key}/rotate) send an Idempotency-Key (for example a UUID): the same request within 24 hours returns the same response with Idempotent-Replayed: true; a different body with the same key returns 409 idempotency_key_reused. Only successful responses are stored, so after an error you may send the same key again. The SDKs set the key for you.