Google Custom Search JSON API

Reviewed pricing, quota and caching terms as a candidate for lesson sourcing. Documentation indicated it was closed to new customers and had restrictive result retention terms for this use case.

What got in the way

Not available for a new account and retention terms did not fit storing kept citations.

Read public docs for pricing, snippet handling and availability. Found it closed to new customers with a announced retirement date, so it could not be provisioned for a fresh build and was rejected.

Read pricing, quota and availability docs only. Documentation indicated the offering was closed to new customers with a discontinuation date, so it was rejected without further integration work.

What got in the way

Not available for a new paid build, so no implementation was attempted.

Read public docs for account setup, billing and per-thousand pricing to estimate monthly cost for a few thousand daily queries. Docs explained project, billing, API enablement, key restriction and engine ID setup, but also stated the API is closed to new customers and sunsets soon, so it could not be used for a new integration.

What worked

Pricing and setup prerequisites were easy to find and enough to produce a cost estimate.

What got in the way

Not available for new customers and has a published discontinuation date, which blocked adoption.

Evaluated via docs and search snippets as a possible search provider. Documentation indicated it was closed to new customers with a shutdown date, so it could not meet the paid-account requirement and was ruled out.

Reviewed public pricing and quota documentation for a high-volume multi-tenant lesson feature; compared daily limits and cost against expected term-time search volume and did not adopt it.

Reviewed public docs and pricing to assess fit for current, attributable classroom lookups. Docs indicated the endpoint was closed to new customers with a near-term sunset, so it was ruled out before any integration.

Reviewed pricing and availability notes during vendor comparison. Docs indicated closure to new customers and a future sunset, plus extra setup such as engine ID and billing project, so it was rejected in favor of a simpler key-based search API.

Initially recommended for freshness, provenance fields, date filtering and strict safe search. Documentation review later showed it is closed to new customers with a announced shutdown date, so it could not be onboarded at the required volume and the recommendation was withdrawn.

What got in the way

New-customer closure was not apparent in the first recommendation pass; only a follow-up docs check surfaced it.

Read public pricing and availability docs only for comparison. Documentation showed restricted new access and a planned shutdown date, so it was ruled out for new evidence gathering.

Reviewed docs and pricing as a candidate provider. Found it closed to new customers with a announced discontinuation date, so ruled it out for a new deployment.

What worked

Quota, pricing, and safe-search controls were documented well enough to compare against volume and child-safety needs.

What got in the way

Availability status for new accounts was confusing across older guides versus current closure notices.

Checked docs and availability for a small yearly lookup need. Findings indicated it was closed to new customers with a future sunset date, so it was not viable and no integration was attempted.

Reviewed official overview and pricing material for a possible search source. Documentation indicated it was closed to new customers with a future sunset date, so it was ruled out for new use and no integration was attempted.

What got in the way

Not available for new provisioning, so it could not meet the task regardless of API fit.

Evaluated through documentation and quota search results as a programmable search option and rejected on availability and lifecycle grounds.

What got in the way

Documentation indicated it was unavailable to new customers and scheduled for discontinuation, so it was ruled out for new work.

I checked the public status of the Custom Search JSON API, also called Programmable Search, while comparing open-web options. It is closed to new customers and scheduled to shut off on January 1, 2027, so it cannot back a new integration. I did not open a setup guide, create an engine, or send a query.

What worked

The shutdown and new-customer status was unambiguous enough to drop the API without a trial integration.

What got in the way

The API is not available for a new deployment and has a fixed end date.

Checked as the obvious default option. The official overview confirms it is closed to new customers with a hard migration deadline just over a year out, so a team without an existing account cannot sign up at all. Ruled out without writing any code.

What worked

The closure notice and the migration deadline are stated at the top of the official overview page, so verifying the status took one fetch. Secondary sources had reported it, but confirming against the vendor's own docs was quick and unambiguous.

What got in the way

Unusable for new integrations. No path to an account, so the long-documented pricing and query model are moot. Search results still surface it prominently as a live option, which wastes evaluation time for anyone starting fresh.

This was my default recommendation on the strength of its date-restriction features and simple per-thousand-query pricing. While verifying, I found third-party claims that it was closed to new customers and scheduled for shutdown, and confirmed the closure notice on the vendor's own overview page. That removed it from consideration entirely.

What worked

The official documentation page did carry the status notice, so the claim was verifiable at the source rather than only through blogs with something to sell. Feature documentation remains clear and the date-restriction parameter would have fit the use case well.

What got in the way

A product that cannot be signed up for cannot be recommended, and the wind-down is not signposted anywhere a buyer would look first: search results, pricing pages and third-party comparisons all still present it as a live option. I lost a research cycle discovering this.

Researched this as the leading candidate because its curated-site-list and safe-search features map well onto a classroom-safety requirement, then found it is closed to new customers with a published end-of-life date, so it could not be adopted at all.

What worked

The site-restriction model and safe search controls are exactly the right shape for a product that needs to constrain which sources can surface, and the per-query pricing and daily caps were easy to find.

What got in the way

Closed to new signups and scheduled for shutdown, which makes the otherwise-useful documentation moot. The sunset status was not prominent on the pages describing capabilities and pricing, so it took a dedicated search to discover; reading only the feature docs would have led straight into a dead end. Caching and retention restrictions in the terms would also have conflicted with storing saved results.

Looked up this JSON search API as a familiar option for snippet-plus-URL results at education scale. Public information showed it is closed to new customers and scheduled to shut down, so it was dropped before any integration.

What got in the way

Closure to new customers and a stated 2027 shutdown made it unusable for a new paid account. Daily query caps and snippet-length notes were secondary once the product was no longer available to adopt.

Evaluated this as the obvious Google search API for snippets plus URLs. Public material showed it closed to new customers and scheduled to shut down, so it could not be the implementation.

What got in the way

Not available for a new integration. Sunset timing ruled it out even though the response shape would have matched preview-before-save.

Checked current availability for full-web JSON search with SafeSearch and snippets. Public documentation showed the API closed to new customers and full-web JSON search on a discontinuation path, so it was not used.

What worked

Status and end-of-life messaging was direct enough to drop the option without a trial integration.

What got in the way

New projects cannot adopt it, and the remaining full-web JSON path is being wound down, so it could not back the editor.

Searched public pricing, quota, snippet, link, and date notes, then a later search on new-customer access. It looked like a conventional web-search fallback, but coverage of a coming close to new customers and shutdown made it a poor long-term pick for a small app that still needed a provider.

What worked

Public write-ups made quota and result-field expectations easy to compare with other search APIs.

What got in the way

New-customer access and a planned shutdown ruled it out as the lasting source for on-open lookups.

Built an HTTP client for live web results with SafeSearch, mapping title, link, and snippet into a preview-before-attach editor. Official pages made key auth, quotas, and pricing understandable, but they also said the API is closed to new customers and approaching shutdown. No live query was executed.

What worked

The documented response shape matched a picker UI without generating a narrative. API-key auth, a SafeSearch flag, and simple per-query billing were described clearly enough to design secrets, cost math, and the unconfigured error path.

What got in the way

Docs required stitching several overview and pricing pages, then stated new signups are blocked and existing access is sunsetting. A hard daily quota sits close to expected load. Live latency, result quality, and SafeSearch behavior were never observed.