Start a bulk product import from an uploaded CSV
- Credentials
- OAuth access token
- Scopes
listings:write- Subscription
- Needs an active plan. Otherwise
403 subscription_locked. - Idempotency
- Send an
Idempotency-Keyheader. A retry with the same key does the work once.
Upload the file first (createProductImportUploadLink, then PUT it) and pass its uploadId. Say how the columns map with a saved profile (profileId, listProductImportProfiles) or an inline mapping; one of the fields mapped has to be a product ID, SKU or barcode, so next month's file finds this month's products instead of creating them again. The file is read and staged here (a file with no rows, a missing column or over 20,000 rows is 422 and nothing starts), then a worker applies it in chunks: this answers 202 at once with the run, status: queued. Poll getProductImport until status is applied (written), previewed (a dryRun: the same run, nothing written) or failed.
Rows with a match are updated (only what the file has a value for: an empty cell never clears a field) and the rest are created as drafts. New products count against the plan's listing cap as in the console: past it, those rows are refused with a reason. Rows that fail are listed in errors (the first 200) and counted in summary.error.
The same uploadId is the same run: asking again returns the run that is there. Needs the Products kit (404 kit_not_found) and answers 501 not_available where bulk import is not switched on.
Access: OAuth scope listings:write; write-gated by the subscription.
merchantOAuth listings:writeAuthorizationBearer <token>OAuth 2.1 (authorization code + PKCE S256, public clients, dynamic registration) for connected AI apps. Effective power is scopes granted AND the person's role today AND the plan. orders:refund, billing:write and team:write are opt-in on the consent screen. Write implies read for marketing:read and messages:read, so a grant holding marketing:write passes a marketing:read check, and messages:send passes messages:read.
Scope: listings:write
storeId*stringA store the person can reach. Any other id answers 404, never 403.
Idempotency-Key*stringRequired on every write. Minted on the device when the person taps, reused on every retry. A write without one, or with one that does not match the pattern, is refused with 422 validation_error. Scoped to the token's subject, remembered 24 hours. A replay returns the first 2xx response with Idempotent-Replay: true; a request still running answers 409 in_flight; the same key with a different method, path or body answers 409 idempotency_key_reused; non-2xx results are not remembered. If the key store cannot be read nothing is processed and the answer is 503 idempotency_unavailable: retry with the same key.
^[A-Za-z0-9_-]{8,128}$application/json- body
Send profileId or mapping, not both.
uploadId*stringprofileId?stringmapping?Your column for each field the import can fill, such as { "sku": "SKU", "title": "Name", "price": "Price" }. At least one of externalId, sku or barcode has to be mapped to start an import.
dryRun?booleanPreview only. The run reads the file and says what it would do; nothing is written.
fileName?stringShown in the import history. Default import.csv.
length <= 200OK
application/json- response
data*curl -X POST 'https://api.shopeazy.co/v1/merchant/stores/<storeId>/product-imports' \ -H 'Authorization: Bearer <access_token>' \ -H 'Idempotency-Key: <unique_key>' \ -H 'Content-Type: application/json' \ -d '{ "uploadId": "string"}'{ "data": { "id": "string", "status": "queued", "dryRun": true, "fileName": "string", "totalRows": 0, "rowsDone": 0, "summary": { "create": 0, "update": 0, "unchanged": 0, "error": 0, "newCategories": [ "string" ] }, "errors": [ { "row": 0, "message": "string" } ], "errorsTruncated": true, "profileId": "string", "failure": "string", "createdAt": 0, "startedAt": 0, "finishedAt": 0 }}Past product imports, newest first GET
Runs started through the API and the console's imports; the newest 200 are listed. Access: OAuth scope `listings:read`; not write-gated. Paged: `{ items, nextCursor }`, `limit` (default 20, at most 50) and an opaque `cursor`.
A signed link to upload the CSV straight to storage POST
Returns a short-lived (10 minutes) signed HTTPS PUT link into the business's upload area, with its `uploadId`. PUT the file's bytes to `uploadUrl` with plain HTTP (for example `curl -T stock.csv -H "Content-Type: text/csv" -H "x-goog-content-length-range: 1,10485760" "<uploadUrl>"`), sending exactly the returned `headers`: the link accepts that content type and 1 byte to 10 MB. No browser CORS is involved. Then pass `{ uploadId }` to startProductImport. A link not used in time stops working, and an unused upload is cleared after a day. Access: OAuth scope `listings:write`; write-gated by the subscription.