2 min read

Google axes Custom Search API on Jan. 1, 2027

Google will shut down the Custom Search JSON API on Jan. 1, 2027. Developers relying on Google web results now have about nine months to migrate.

Image: Hacker News

Google’s Custom Search JSON API will return HTTP 410 Gone on January 1, 2027, ending a long-running tool many developers used to get Google web search results as JSON. According to the source, the shutdown notice appeared in January 2026 in a low-profile Google Developers blog post and in console email notifications, and the API was still technically online as of August 2026.

That matters because Google’s suggested alternative, Vertex AI Search, is not a direct replacement. It is designed for searching your own corpus, not the public web, with pricing starting at about $2 per 1,000 queries for a basic tier and rising to $7+ with extensions. The source says Google’s guidance for public web search at scale is to “explore third-party providers,” effectively signaling an exit from this business.

For years, the API served three main use cases:

  • Site search on a company’s own website
  • Web search for bots, assistants, and RAG pipelines
  • Research and academic workflows, including citation checks and corpus building

The source cites Google aggregate stats saying observed usage breaks down to roughly 40% site search, 35% bot or assistant search, and 25% research and academic work.

Recommended reading

LLM coding feels faster — and more exhausting

Third-party alternatives including SerpApi, ScaleSerp, and Bright Data SERP API can provide Google results, but each uses a different response format. That means teams moving off CSE usually have to rewrite their parsing logic. The article’s main pitch is a separate tool, google-cse-replacement on Apify, which it says reproduces the exact CSE v1 JSON schema so existing code that depends on fields like searchInformation.totalResults, items[].title, items[].link, and items[].snippet can keep working with minimal changes.

The pricing comparison in the source for 100,000 searches per month looks like this:

  • Google CSE: $500
  • Vertex AI Search: $200–$700+
  • SerpApi: $500
  • ScaleSerp: $75–$100
  • Bright Data SERP: $500
  • google-cse-replacement: $500 at $0.005/query

The source also highlights quirks in CSE’s current quota model:

  • 100 queries/day on the free tier
  • $5 per 1,000 queries on paid usage
  • Up to 10,000 queries/day per engine ID
  • Teams often used multiple engine IDs to get beyond that limit
  • Sustained traffic above 50 QPS could trigger throttling

That workaround disappears when the API shuts down. The replacement described in the source reportedly has no per-engine cap and can sustain roughly 100–200 QPS, with 500+ QPS bursts, depending on Apify concurrency limits.

The source frames migration as a relatively small code change for teams that only need CSE-compatible output. But the bigger takeaway is simpler: after nearly two decades of serving as an officially sanctioned way to query Google’s public web index, Custom Search is going away, and developers depending on it have until January 1, 2027 to move.

Tomas Berg

Computing Editor

Tomas lives in the terminal. He covers chips, laptops, and operating systems with a focus on performance and efficiency. He reads kernel changelogs the way other people read fiction, and he's always on the hunt for the perfect mechanical keyboard switch. If it processes data, Tomas has an opinion on it.

via Hacker News

// Keep reading