Separate link discovery data from ordering operations
A backlink data API exposes observations from a provider’s index in a structured format. An ordering API instead creates or retrieves commercial orders. Receiving rows from an index does not publish links, and an order receipt is not an all-web discovery result. Decide which job your integration must perform before choosing credentials, data models or billing assumptions.
The example here uses the documented DataForSEO v3 Backlinks Live shape. The on-page inspector reads a saved response locally; it does not connect to that provider, request credentials or call a paid endpoint. The fixture is authored and fictional, with reserved example domains. It illustrates a format, not an actual scan or a claim about any publisher.
Read the envelope before reading the rows
The endpoint documentation distinguishes response status, task status, request data and result items. The importer requires one successful task and one result from the specified endpoint. A response error or task error must not become zero backlinks. Check the provider’s error definitions rather than treating an HTTP response alone as application success.
Preserve target, grouping and filters with the records. A domain target, page target and one-per-domain grouping answer different questions. The provider documents consolidation of references from the same source page; returned item count therefore cannot be assumed to count every physical anchor. Provider totals describe the request’s scope and index, not an independently verified whole-web inventory.
Read the fictional response as four observations
The example declares nine matching results but supplies four items. Five are absent from this file, not proved lost. One record has an ordinary provider follow flag, another has sponsored and nofollow attributes, a third is marked lost by the provider, and the fourth contains unknown values. The last-seen dates are copied provider observations, not fresh checks made by this page.
The inspector keeps false, zero and missing values distinct. It retains repeated source–target pairs if they appear, rather than silently combining observations. The table explains six interpretation decisions. An API-shaped JSON file can be authored by anyone, so parsing it validates the supported structure; it does not authenticate the origin or prove that its claims are correct.
| Check | Fictional value or field | Interpretation | Do not infer |
|---|---|---|---|
| Envelope and task | status_code: 20000 at both levels | Supported successful response structure | Fresh source-page checks |
| Total versus returned | total_count: 9; items_count: 4 | Four supplied records; partial result | Five links lost or all links known |
| Source and target | url_from; url_to | Preserve the exact reported URL pair | That one item equals every anchor on a page |
| Typed relationship values | dofollow: false versus null | Reported false and unknown remain different | Missing means follow or nofollow |
| Historical observation | is_lost: true; last_seen: 2026-10-03 | A provider report of loss and its observation date | A new deletion check today |
| Request context | mode: as_is; backlinks_status_type: all | Keep selected scope alongside records | That another filter or grouping has the same total |
Plan a bounded integration before spending API units
A real client needs server-side credentials, an approved request budget and a record of request parameters. Review current provider limits and pricing before dispatch. Save the response before transforming it. Pagination must preserve the relevant query scope; compare the number actually received with the provider’s total and follow its documented continuation rules. Do not automatically retry an uncertain paid request as if it were free.
Keep unknowns and provider-specific definitions in your schema. A lost flag is a provider observation; it is not a new deletion check. A boolean follow field does not preserve every relationship or page-level instruction, so retain reported attributes as well. Google’s Search Console API reference describes different services and does not provide a Links report method. Search traffic rows are not substitute backlink records.
Use the local importer and keep the original
Load the fictional example, paste a supported saved response or select a UTF-8 JSON file. Review the reported total, actual row count, partial-result notice and request context. Export CSV for tabular work and projection JSON for typed values. Missing and empty text both become blank CSV cells; JSON preserves their distinction. Spreadsheet-like text receives a protective apostrophe in CSV, so keep the original JSON for exact source values.
For offline use, download the importer module, command-line wrapper and language-specific example into the same folder. The commands below read a local file with Node.js and write CSV or JSON; they make no network call. The module is generated from the same tested source as the browser tool. The worked and blank CSV sheets below are interpretation notes, not inputs to the JSON importer.
Download the worked API review · Download a blank API review
node backlink-api-cli.mjs backlink-api-example-en.json > backlink-api-records.csv
node backlink-api-cli.mjs backlink-api-example-en.json --json > backlink-api-projection.jsonKnow the limits of the projection
This tool accepts one task, one result, at most 1,000 items and 512 KiB. Selected request metadata is bounded to 2,000 characters. It exports selected fields, not every provider field; keep the raw response. Dates are retained as supplied rather than converted into publication dates. It does not fetch source URLs, merge paginated files or authenticate an account. An empty valid response remains scoped to that response.
Use the export as an input to a review process, then obtain current page evidence where needed. Attribute delivery, source availability, provider history, search-engine indexing and ranking outcomes remain separate. This small importer is a reproducible integration example, not a production SDK or a promise of complete backlink discovery. Extend it only after defining the additional provider contract, limits and tests.