Google’s full-web Google Search API now has its most detailed public documentation to date. On September 9, 2026, Google refreshed its developer pages for the Web Search Service API, the partner-only service that lets approved companies retrieve and display Google Search results in their apps and websites. The update clarifies exactly what a request must include to be accepted. What it does not clarify: who qualifies, what it costs, and whether existing Custom Search JSON API users can migrate before the January 1, 2027 shutdown.
How Does Google’s Partner Search API Work?
The Web Search Service API lets “programmatic partners” (Google’s term for approved business customers) retrieve Google Search results in JSON format (a standard data exchange format) via REST or gRPC (two common ways apps communicate with servers over the internet). According to the updated developer documentation, every request must include three components: an active Google Cloud project, an API key generated through Google Cloud Console, and a partner client ID tied to a signed partner agreement.
The API supports one primary method, called Search, which performs a full web search across Google’s entire index. Responses return up to 20 results per request (10 by default), with additional pages available via a pagination token. Partners can filter results by language, country, and date range, enable SafeSearch, or sort by date. Each result includes the title, URL, snippet, MIME type (file category, such as PDF or HTML), file format, an estimated total result count, and a corrected query if Google suggests one. Including the user’s IP address in each request helps with regional routing and misuse prevention, according to the updated invoke guide.
Why Does the January 2027 Deadline Matter?
In January, Google announced it was winding down two products. The Programmable Search Element, a site-search widget, would be narrowed to cover no more than 50 domains. The Custom Search JSON API, the developer tool many teams use to query Google Search programmatically, would be discontinued on January 1, 2027, and is already closed to new customers. Google directed anyone needing full-web results to fill out an interest form for a separate “full web search solution,” without naming it explicitly.
The September 9, 2026 documentation update connects those dots. The Web Search Service API performs a full web search and requires a partner agreement, matching both criteria Google described in January. As of September 11, 2026, according to Search Engine Journal, neither the January announcement nor the updated documentation links to the other.
What Information Is Still Missing?
The updated pages are more detailed than anything Google has previously published on this service, but they leave three critical questions unanswered.
- Eligibility: The documentation does not explain how a company becomes a “programmatic partner” or what requirements apply. The January interest form is still the only public entry point.
- Pricing: No cost information appears in the documentation or the January announcement.
- Query limits: The documentation confirms a maximum of 20 results per request but does not state daily or monthly volume caps.
For developers currently using the Custom Search JSON API for full-web projects, Google has not confirmed whether they can migrate to the Web Search Service API or what steps that transition would require.
Where Does This Leave Teams Using Custom Search?
Google has pointed users in two directions depending on use case. For enterprise needs like conversational search and grounding (connecting an AI model to live web data to improve the accuracy of its responses), Google pointed users toward Vertex AI Search. For full-web programmatic access at scale, the Web Search Service API appears to be the intended path, but access requires a partner agreement with no public application process beyond the January interest form.
Search Engine Journal notes that Google’s January post, which first described the full-web solution, still directs users to an interest form that does not explain partner requirements. The updated Web Search Service documentation does not reference that form or the January announcement at all.
Before Your Team Makes a Move: A Quick Checklist
- Confirm whether your integration uses the Custom Search JSON API and whether it queries more than 50 domains or the full web.
- Submit Google’s interest form from January if you have not done so yet. Partner agreements likely require meaningful lead time to process.
- Evaluate Vertex AI Search as an alternative if your use case centers on AI-powered or conversational search, or grounding.
- Review your application’s requirements against the Web Search Service API’s documented parameters: language and country filtering, SafeSearch, date-based sorting, and pagination.
- Set an internal review deadline well before January 1, 2027, to allow time for any contract negotiation a partner agreement might require.
- Monitor Google’s developer blog and the Web Search Service documentation page for updates on eligibility and pricing.
What This Means for AI-Search Visibility
The clearest takeaway for any brand thinking about AI-search visibility: programmatic access to Google’s full index is becoming a gated, partner-level arrangement, not an open API anyone can activate.
Think of it this way. Until now, a developer could enable the Custom Search JSON API in a Google Cloud project and start pulling Google results into almost any product, including AI-powered search tools, chatbot grounding systems, or content discovery features. From 2027, doing that at full-web scale requires a negotiated agreement with Google. The bar is higher, and the process is opaque.
This matters for AI agents (automated software that queries information on a user’s behalf to answer questions or complete tasks) that currently depend on a programmatic Google Search API to find and surface sources. If those agents rely on the Web Search Service API going forward, the companies building them will need partner status. Smaller AI tool developers may be redirected toward Vertex AI Search instead, which operates through a different mechanism and may surface content differently from standard Google Search.
One implication the sources do not address directly: content surfaced through the Web Search Service API will be drawn from Google’s full index, weighted by the same ranking signals as standard organic Google Search. Brands with strong search presence, good indexability, and solid domain authority are positioned to appear in whatever AI-powered products partner tools deliver. Brands that only rank in lower-authority or narrow-niche positions may find themselves underrepresented. This is a reason why standard SEO fundamentals remain the most durable investment a brand can make ahead of this transition, separate from any AI-specific strategy.
Hingewise’s read: the January 2027 transition is as much an infrastructure decision as a product change. How Google structures partner-level access to its index will determine which AI tools can credibly serve full-web results, and therefore which brands surface when those tools respond to user queries. That story is moving slowly right now, but it will accelerate as partner agreements start to be referenced publicly or disclosed by AI product companies.
The next signal to watch: whether Google publicly names the Web Search Service API as the official migration target for Custom Search JSON API users, and whether it opens a clearer application path for mid-size developers who are not yet “programmatic partners.”
Sources: Search Engine Journal, September 2026; Google Web Search Service API documentation, updated September 9, 2026.
This article was drafted with AI assistance from the sources cited above and checked against them before publishing. The "What this means for AI-search visibility" section is Hingewise's editorial analysis, built on the frameworks and positions we use in our GEO work. How we work: Editorial policy.