**The whole search page, not just ten blue links.** Most SERP APIs hand you a title, a URL and a snippet. This one returns everything Google actually showed — **AI Overviews, Knowledge Graph, sitelinks, People Also Ask, related searches, perspectives and spelling corrections** — in one call, on one credit.
And it is **one key for every engine**. Google, Bing, DuckDuckGo, Yandex, Baidu, YouTube, News, Images, Maps, Shopping, Scholar, Jobs, Trends, Amazon, eBay, Walmart — **40+ engines behind a single endpoint**. Change the `engine` parameter and you are searching somewhere else. No second subscription, no second integration, no second bill.
## 💸 The same depth of data, at a fraction of the cost
Structured knowledge graph, AI overviews and sitelinks are normally sold as a dedicated SERP data plan — fixed tiers, bought whole, and sized well above what most projects actually need. Want 6,000 searches a month? The smallest plan that covers it holds 15,000, and you pay for all of them.
| Your monthly volume | Dedicated plan you would need | That costs | **This API** | **You save** |
|---|---|---|---|---|
| 6,000 searches | 15,000-search tier | $150 | **$25** | **83%** |
| 25,000 searches | 30,000-search tier | $275 | **$80** | **71%** |
| 70,000 searches | 100,000-search tier | $725 | **$175** | **76%** |
Per search that is **$0.025 → $0.0042**, **$0.011 → $0.0032** and **$0.0104 → $0.0025**. Pay for what you use, month to month, cancel anytime — no annual commitment and no minimum.
*Compared against published list pricing for dedicated SERP data plans returning equivalent response depth.*
## 🔍 What comes back in a single response
A search for a brand, product or person returns, when Google shows them:
🤖 **AI Overviews** — Google's generated answer, as structured text blocks with the source references behind it. Not a scrape of the box; the blocks, headings, lists and links.
🧠 **Knowledge Graph** — the entity panel. Type, description, images, and the structured attributes: for a book series, the titles; for a company, the founders and headquarters.
🔗 **Sitelinks** — the expanded sub-links under a result, each with its own title, URL and snippet.
❓ **People Also Ask** — the questions, with the tokens to expand each one.
🔎 **Related searches, perspectives and "things to know"** — what else the page offered.
✍️ **Spelling metadata** — the query as typed, the query Google actually used, and the correction itself as its own field. You can always tell whether a correction happened.
📄 **Organic results** with position, title, link, displayed link, snippet, highlighted terms, source and favicon.
## Why this API
🔑 **40+ engines, one key** — no second subscription when you need images, news, Maps or Scholar. Most providers sell those as separate APIs, each with its own plan and its own integration.
📊 **The full page, not a summary** — knowledge graph, AI overviews and sitelinks are returned as structured JSON, not left for you to scrape.
🔁 **Already built on SerpApi? Change one line.** Responses use the **SerpApi schema**, field for field — `search_metadata`, `search_parameters`, `organic_results`, `knowledge_graph`, `answer_box`, `related_questions`. Point your base URL here and your parsing code keeps working.
📄 **Pagination that works** — every response carries `api_pagination` with a ready-made link to the next page. Follow it, or set `start` yourself. No tokens to decode.
🛡️ **Licensed data, not an anonymous scrape** — the results carry a legal shield, which matters if anyone in your organisation ever has to sign off on where your data came from. Scraper fleets get blocked, rate-limited and occasionally disappear; this does not.
🔁 **Automatic failover** — requests are routed across a pool of capacity with health checks and circuit breaking, and retried automatically when one path fails. You get resilience you did not have to write.
🩺 **A free health check** — `GET /ping` tells you whether the service is up. No key needed, no quota consumed, poll it as often as you like.
📊 **Traceable** — every response carries `X-Request-ID`. Quote it and a failure becomes one row in our logs.
🆕 **New engines work immediately** — parameters are forwarded verbatim rather than enumerated, so a newly supported engine is available here the same day, with no version bump and nothing to wait for.
## 🌍 Targeting
| Parameter | What it does |
|---|---|
| `gl` | Country — `us`, `gb`, `in`, `de` |
| `hl` | Language — `en`, `es`, `hi` |
| `location` | City-level, e.g. `Austin, Texas, United States` — what you want for local rank tracking |
| `google_domain` | `google.co.uk`, `google.co.in` |
| `num` · `start` | Results per page, and pagination |
Every Google advanced operator works inside `q`: `site:`, `inurl:`, `intitle:`, `filetype:`, quoted phrases and boolean terms.
## 📏 Limits — read these before you build
| Limit | Value | If you exceed it |
|---|---|---|
| Requests per second | Set by your plan | `429` — retry with backoff, honour `Retry-After` |
| Requests per month | Set by your plan | `429` — upgrade, or overage applies on ULTRA and MEGA |
| Response size | ~40–60 KB for a rich Google query | Counts toward your RapidAPI bandwidth allowance |
| Result freshness | 10 min to 24 h, by engine | Send `no_cache=1` to force a freshly fetched result |
Freshness windows are set per engine, because how long a result stays true is a property of the engine and not of this service: news and trends refresh every 10–15 minutes, general web search hourly, Maps every six hours, Scholar daily.
## Straight answers about what we don't do
We would rather you know now than find out after subscribing.
⚡ **A live search takes a few seconds.** Typically 5–7 seconds for a fresh Google query. Normal for this category, but slower than an API that only returns ten links. If you need sub-second on every call, this is not the right tool.
📦 **One query per request.** Batching several queries into a single call is on the roadmap, not shipped.
🔄 **No realtime streaming.** Every call is request/response.
🗺️ **Engine coverage tracks the search engines themselves.** Parameters are forwarded rather than enumerated, so a new engine works the day it appears — and one that is retired stops working here too.
**Perfect for:** SEO monitoring and rank tracking, keyword research, competitive analysis, market research, LLM and RAG pipelines needing live web grounding, price and product monitoring, and anything that has to search more than one engine.