The details, without the ambiguity

No. In the native app, titles are classified and searches run entirely on-platform in Apex. There are no external callouts and no data is sent away to be enriched. The standalone API is a separate, optional product for systems outside Salesforce.

Yes. Classification is pattern-based, not model-based. The same title with the same pattern version always produces exactly the same result, which makes results safe to cache and audit.

Searches execute in user mode, so field-level security, CRUD, and sharing rules apply automatically. People see only the records they are already permitted to see, and every classification and match is an auditable Salesforce record.

Logic grows deliberately, edge case by edge case. Every classification carries taxonomy and pattern version stamps, so a change is visible and reproducible. A pattern bump means a title may classify differently than before, so you can re-run classifications when you depend on consistency.

A free trial key allows 100 single-title requests per day and cannot use the batch endpoint. A Standard key allows 10 requests per second with a burst of 20, and 100,000 requests per day. A batch of up to 500 titles counts as one request.

A valid response with null or empty classification fields. You always get the full, stable response shape, so unrecognised titles are easy to detect and handle.

ESInsight ships as metadata through a package or SFDX, with its own permission set. Install into a sandbox, validate, and promote it like any other release — no data migration required.