1. Start with the question
Open the Job Scraper preview. The default example searches for “python” in Canada. Changing the query helps you inspect the request shape; it does not collect matching jobs.
2. Make the scope explicit
The request names its sources, country and total result limit. In the planned service, a source must support the requested filters. Sources shown in the example are candidates that still require validation.
3. Decide whether partial is useful
Choose the partial scenario. Fourteen illustrative records were delivered, one source was excluded and another failed. The sample request allows partial delivery. Your own workflow may require all-or-nothing behavior instead.
4. Compare failure with no matches
Switch to the successful, no-matches example, then to failed. Both deliver zero records. Only the former says the selected scope was actually checked. Retrying and reporting should depend on that difference.
5. Read the spending boundary
The example sets maxCostUsd to 0.10. This is a ceiling, not the amount automatically charged. A reservation protects the budget; settlement follows the billable quantity under the published pricing rules.
6. Take the contract into your workflow
The preview documentation explains the proposed input and outcome semantics. Exact production endpoints, supported sources, authentication and final rates will be published when access opens. Until then, the examples are a design walkthrough.