CYBEXO Pulse Public Scan API
Send requests to https://pulse.cybexo.com. The compatibility host is https://api.cybexo.io. A successful creation request completes a source scan and returns the report in the HTTP response. The public API operates independently of account enrollment and alert delivery. Download the OpenAPI document
Endpoints
Section titled “Endpoints”| Method and path | Success response |
|---|---|
GET /v2/public-scan/healthz |
HTTP 200 with public scanner service status. |
POST /v2/public-scan/scan |
HTTP 200 with data.scan and data.report. Use this path for new integrations. |
POST /v2/public-scan/jobs |
Compatibility alias for the same synchronous source scan. Takes the same input and returns the same HTTP 200 envelope. |
GET /v2/public-scan/report/{token} |
HTTP 200 with a saved report. A historical browser-job token may still yield HTTP 202 with data.job. |
Create a website scan
Section titled “Create a website scan”Send a bare hostname you own or are authorized to assess. The service discovers public pages from the start page, links and sitemaps. The default and deployed cap are 300 pages. Request a smaller positive integer with maxPages. Larger requests are capped at 300. The total download budget is 100 MiB. Fetch, robots, discovery and execution limits may finish the scan earlier.
curl -sS https://pulse.cybexo.com/v2/public-scan/scan \ -H 'Content-Type: application/json' \ -d '{"domain":"example.com","maxPages":10}'
Check an exact page
Section titled “Check an exact page”Pass a complete public HTTPS URL in domain. Alternatively, send a bare hostname in domain and its same-host HTTPS landing-page URL in target_url. An exact request evaluates one page without following its links or expanding its sitemap. Pulse can also read canonical CYBEXO loader metadata where applicable.
curl -sS https://pulse.cybexo.com/v2/public-scan/jobs \ -H 'Content-Type: application/json' \ -d '{"domain":"example.com","target_url":"https://example.com/offers/"}'Submit an anonymous public URL and omit credentials, custom ports, fragments, private query parameters and signed values. Both endpoints also accept domain and maxPages as query parameters. JSON body values take precedence. Use the ordinary source request without a profiles field.
Read the creation response
Section titled “Read the creation response”The envelope contains ok, data and meta. On success, read data.scan immediately. Save data.report.token, url, json_url and expiresAt. The human-readable link is absolute. Resolve the API path in json_url against the service origin. Repeating a successful POST creates a fresh report.
| Field | How to interpret it |
|---|---|
scan.scan_method, source_standard, source_target |
Current values are source_html, pulse-source-v1 and either page or site. |
scan.assessment |
A supported source failure yields fail even when other pages are missing. A complete, clean source scan yields pass. Incomplete evidence without a confirmed failure yields inconclusive. |
scan.pages[].url, scan.pages[].page_id |
The display URL masks query values and any path segment containing @ or %40, or longer than 80 characters in its encoded form. Other path segments and query names remain visible. Distinguish pages with the report-local page_id when their display URLs match. |
scan.pages[].source_verdict |
pass or fail for evaluated source, unreadable for a page Pulse could not assess, or not_scanned for a URL left unchecked by the byte budget. |
scan.pages[].check_results |
When a readable page has recorded rule results, read each check’s key, status and evidence source. An explicit unknown for google_consent_source_order means the order could not be determined. Older reports can omit these results. |
source_issues, source_notes, source_cmp on a page |
A source issue gives a code and detail, with evidence or a fix when available. CMP signals and notes describe what was visible in the HTML. |
scan.coverage |
Read discovered, attempted, tested, failed, excluded, complete and limit_reached. The failed count concerns page retrieval. Read page verdicts for source findings. A total-byte stop can also provide not_scanned and limit_reason: "total_bytes". |
A recognized source_cmp can carry registry_version and registry_updated_at for that page. A new scan includes google_web and iab_tcf only when the snapshot is available and inside its seven-day window. An omitted flag means unknown. For Google, false can indicate a listing for another platform or an absent provider. For IAB, it can indicate deletion or absence. For a historical report, use the original scan time and recorded reference date. If either date or the reference identity is missing, listing freshness is unknown. Source rule catalog
Coverage applies to the chosen scope. Treat missing fields in old reports as unavailable. A source pass can coexist with undetermined Google tag order. Confirm browser execution separately. Scanner guide
Retrieve and share a report
Section titled “Retrieve and share a report”Use the token exactly as supplied. New reports carry data.client.domain, data.scan, data.generatedAt and data.expiresAt on retrieval. Save the link only for as long as needed. Anyone with it can read the report until expiry. Older monitoring reports can omit timestamps or use different retention.
Reopening an older public link applies the current URL masking to the displayed pages and the CSV and JSON exports. The stored observation time and verdict keep their original values. Previously downloaded files retain their earlier contents. Keep the original anonymous target separately if you need to rescan an exact URL with a query or masked path.
curl -sS "https://pulse.cybexo.com/v2/public-scan/report/${REPORT_TOKEN}"Migrate clients that expect a job
Section titled “Migrate clients that expect a job”Previous OpenAPI versions described new POST /jobs calls as asynchronous browser jobs. Update generated clients to use the current specification. Handle the completed HTTP 200 result on both creation paths and read data.scan directly. A response from either POST contains the saved report reference. Clients can then use GET whenever they need the saved result.
Keep historical GET handling for existing tokens. For a 202 read, inspect data.job.status. Stop on a terminal failure. For a pending job or publication still in progress, retry the same GET after the supplied Retry-After delay when present, with a bounded wait. Historical reports may include browser or verification fields that describe observations made on the recorded run.
Handle errors
Section titled “Handle errors”| HTTP status | Action |
|---|---|
| 400 | Correct invalid input, such as a malformed token, invalid JSON, an unsuitable URL or a nonpositive page limit. |
| 404 on GET | The report is missing or expired. Preserve the original public input if a fresh scan will be needed. |
| 408 or 413 on POST | Send a complete request body within five seconds and keep it within 16 KiB. |
| 422 | An explicit verification profile is unsupported. Use the ordinary source request if it meets your needs. |
| 429 | Rate limit reached. Observe Retry-After before requesting another scan. |
| 503 | Read error.code. Admission may be temporarily unavailable or an explicit verification mode may be unavailable. Respect Retry-After when provided. |
| 500 | Processing failed. Capture meta.request_id for support and retry deliberately. |
A network timeout leaves the creation outcome uncertain. Reuse a report link if you received one. Sending a new POST can create another report. Report privacy and removal