Skip to content
Start free Sign in

Kukie.io API changelog

Changes to version 1 of the Kukie.io API, newest first.

Each entry is a version of the API's OpenAPI document, which the API reference is built from. Additions can arrive at any time. A change that could break your software is announced here before it takes effect, as section 8 of the API Terms describes.

1.9.0 -

  • A new webhook event type, report.ready, says that the scan report of a completed scan is ready. Its data names the site and the scan, gives the report's schema_version, content_hash, built_at and pdf.status, and carries in links the addresses of the JSON report and the PDF, which need an API key with reports:read. An endpoint can subscribe to it now. Scan reports are not switched on yet, so no report.ready event is sent until they are announced here. You do not have to change anything.
  • The description of GET /v1/sites/{siteKey}/scans now says that old scans are deleted every day, not once a week, and that a deleted scan's report goes with it. You do not have to change anything.

1.8.0 -

  • A new operation, GET /v1/sites/{siteKey}/scans/{scanId}/report/pdf, returns the scan report as a PDF file, printed from the same document as the JSON report, with the permission reports:read. A PDF that is still being made answers 409 report_not_ready with a Retry-After header, and when 10 of your organisation's PDFs are already waiting to be made, the answer is 429 queue_full with Retry-After. Scan reports are not switched on yet, so the operation answers 404 not_found until they are announced here. You do not have to change anything.

1.7.1 -

  • The description of GET /v1/sites/{siteKey}/scans/{scanId}/report now says how many report builds your organisation's requests start: up to 20 in each window of 60 seconds, and one more when two of your requests arrive at the same moment. Scan reports are not switched on yet. You do not have to change anything.

1.7.0 -

  • Two new operations for scan reports: GET /v1/sites/{siteKey}/scans/{scanId}/report returns the scan report of a scan that ran to its end, with the new permission reports:read, and GET /v1/schemas/scan-report-1.json returns the report's JSON Schema without an API key. A report that is not ready yet answers 409 with the new error code report_not_ready and a Retry-After header. API keys created before this change do not have reports:read: create a new key to get it. Scan reports are not switched on yet, so both operations answer 404 not_found until they are announced here. You do not have to change anything.

1.6.0 -

  • GET /v1/sites/{siteKey}/scans/{scanId} has a new member, before_consent: what the scan's first pages set and loaded when the scanner loaded them again as a first-time visitor who has not made a choice in any cookie banner, as a status (not_checked, partial or checked) and counts of the pages, cookies, storage keys and hosts. This check is not switched on yet, so every scan reads not_checked with no page planned until then. The scans list and the answer to starting a scan do not change. You do not have to change anything.

1.5.0 -

  • Webhooks are switched on. Webhook endpoints (GET and POST /v1/webhook-endpoints, GET, PATCH and DELETE /v1/webhook-endpoints/{endpointId}, POST /v1/webhook-endpoints/{endpointId}/rotate-secret and POST /v1/webhook-endpoints/{endpointId}/test) and events (GET /v1/events and GET /v1/events/{eventId}) now answer, with the permissions webhooks:read and webhooks:write, on plans that include the API. To receive and verify events, read Receive webhooks in the Help Centre. You do not have to change anything.
  • The description of DELETE /v1/webhook-endpoints/{endpointId} is corrected: the answer does not wait for what is already on its way, so a message sent before the delete can still reach the endpoint after the answer. Nothing else is sent to it.

1.4.0 -

  • POST /v1/webhook-endpoints and a PATCH /v1/webhook-endpoints/{endpointId} that changes url can answer 429 concurrency_limited with a Retry-After header while another address check or test event of your organisation is running, or while Kukie.io is checking as many addresses and sending as many test events as it allows at once, and 503 service_unavailable when the check can't run because of a fault in Kukie.io. Neither answer writes anything: wait the seconds in Retry-After, then send the request again. A PATCH that sends the url the endpoint already has does not check it again. POST /v1/webhook-endpoints/{endpointId}/test also answers 429 while an address check of your organisation is running. Webhooks are not switched on yet, so these operations still answer 404 not_found.

1.3.1 -

  • DELETE /v1/webhook-endpoints/{endpointId} no longer answers 403 plan_upgrade_required: you can delete a webhook endpoint when your plan no longer includes the API. Every other webhook operation still needs the plan. Webhooks are not switched on yet, so this operation still answers 404 not_found. You do not have to change anything.

1.3.0 -

  • New, and not switched on yet: webhook endpoints (GET and POST /v1/webhook-endpoints, GET, PATCH and DELETE /v1/webhook-endpoints/{endpointId}, POST /v1/webhook-endpoints/{endpointId}/rotate-secret and POST /v1/webhook-endpoints/{endpointId}/test) and events (GET /v1/events and GET /v1/events/{eventId}), which need the permissions webhooks:read and webhooks:write. Until webhooks are announced here, these operations answer 404 not_found. You do not have to change anything.

1.2.1 -

  • Descriptions corrected, with no change to any operation or member, or to the status or code of any answer. Deleting a site names the security email that the owner and every admin get. Creating and deleting a site, and a copy of settings, say how they count toward the budget for banner rebuilds, which now names its limit for a minute. The introduction says when a stored idempotent answer is deleted, when a retry runs again, and that an operation on something with no etag yet must be sent without If-Match.
  • The errors member of an error is also filled with a 409 conflict about a site's domain or external_id, as the API already answered. existing_site_id comes with that conflict when your API key can read the other site.
  • The usual detail of 412 precondition_failed and of 503 service_unavailable gives better advice. Decide by code, never by detail.

1.2.0 -

  • GET /v1/usage can answer 503 service_unavailable with a Retry-After header when its figures take too long to compute. Wait that long, then ask for a shorter window.

1.1.1 -

  • The message of a scan with failed pages gives the number of pages that failed, without their addresses. The addresses stay in the scan's failed_pages.

1.1.0 -

  • A copy of settings (copy_settings_from when you create a site, and POST /v1/sites/{siteKey}/copy-settings) carries the banner settings, the texts and the region rules, and a copy to other sites also, on request, the cookies added by hand and your own services. It never copies the consent logging switch or the scan frequency, and it copies the accessibility widget only when the API key also has the accessibility:write permission. A site left with a widget that differs from the source's gets the warning accessibility_widget_not_copied.
  • GET /v1/account counts only the sites an API key reaches when the key is limited to chosen sites, and its can_create_site is false for such a key.

1.0.0 -

  • The first version of the API: sites, banner settings and texts, region rules, cookies and categories, the accessibility widget, cookie scans, consent records and statistics, the account and API usage.

Listed on