SERP API questions, answered

10,000 searches cost $7.50 ($0.75/1K searches), 100,000 searches cost $56.25 ($0.56/1K searches) and 1,000,000 searches cost $375 ($0.38/1K searches), as of October 2026. The rate falls to $0.30/1K searches with 5,000,000 searches. All seven packs are on the pricing page.

1 credit = 1 successful search request, which returns one page of results on any supported engine. A request that fails is not charged. In practice, one credit equals one HTTP request to https://serpcrypto.com/api/v1/search that returns a results page. A request for page 3 of the same query is a separate credit. Checking your balance with GET https://serpcrypto.com/api/v1/credits costs nothing.

No. Prepaid search credits never expire: unused searches stay on your account until you use them. That is the main difference from monthly SERP API plans, where unused searches usually reset at the end of each billing cycle. A seasonal project or an agent with irregular traffic keeps every credit it paid for.

No free tier. No free tier: test with the smallest pack, paid from the minimum top-up; the part added only to reach the minimum is refundable on request within 30 days, in USDT or USDC, if unspent. The smallest pack is 10,000 searches for $7.50, enough to test every engine on your real queries. Unused credits from that test stay on your account.

Yes. A search on Google, Google News, Google Images, Google Maps, Google Shopping or Bing costs one credit each. You do not need to budget a different rate for maps or shopping results, and a mixed workload is priced the same way as a single-engine one.

No. Every error response is free: invalid parameters (400), a wrong key (401), no credits (402), too many parallel requests (429), a failed search (502) and a timeout (504). Only a successful results page uses a credit. The error reference explains each code and how to retry.

Because Google stopped honouring the num=100 parameter in September 2025. About 10 organic results per page, as Google has served since it stopped honouring num=100 in September 2025; each extra page is one more credit. To collect the top 100 for a keyword, request page=1 to page=10; tracking only the pages you need keeps the cost down.

The minimum top-up is $25. You can then buy any pack from your balance. If you buy 10,000 searches for $7.50, the rest of the top-up stays as balance. If part of the top-up was added only to meet the minimum and stays unspent, you can ask for it back within 30 days, in USDT or USDC.

Google Search, Google News, Google Images, Google Maps (local results), Google Shopping and Bing. You pick one with the engine parameter: google, google_news, google_images, google_maps, google_shopping or bing. The response keeps the same envelope for all of them, so one parser handles every engine. See the search reference.

Yes, when Google shows one on that page. The ai_overview object holds the text and the list of sources it cites, with title and link. If Google does not show an AI Overview for the query, location and device you asked for, the key is simply absent. See the Google Search API page.

Yes. Country (gl), language (hl), Google domain and a city-level location string, on desktop or mobile. Send location=Austin, Texas, United States together with gl=us and hl=en, for example. Local packs and map results depend heavily on the city, so set it for any local SEO work. The locations guide lists the formats.

Yes. Add device=mobile. Google and Bing lay out mobile pages differently: blocks appear in another order and some features show only on phones. Rank trackers usually record both devices for important keywords, which costs one credit per device.

Parsed JSON: organic results, ads, People Also Ask, related searches, top stories, knowledge panel, local pack, shopping and video results, and the AI Overview text with its sources when the page shows one. Every response starts with search_metadata (ID, status, engine, credits used) and search_parameters. The response format page documents every block, field by field, with an example.

For web results, yes. Microsoft retired the Bing Search APIs in August 2025. With SerpCrypto you send engine=bing and receive the organic results and related searches as JSON, paid per search with credits that never expire. Our Bing Search API replacement page compares the options.

It can be, if you need full Google results rather than a custom search engine over chosen sites. SerpCrypto returns the regular results page, with ads, People Also Ask and local results. Status, limits and prices of Google's own API are covered in our comparison.

Yes, with engine=google_maps. Each item of local_results has the name, address, phone, website, rating, review count, category, opening hours and coordinates when Google shows them. Results depend on the location you send. See the Google Maps results API.

No. SerpCrypto, a no-KYC SERP API, asks for no ID, no selfie, no company documents and no company email. Account-number login; email is optional. You still accept the Acceptable Use Policy, and abuse leads to suspension. More on the no-KYC page.

BTC, USDT, USDC, ETH, SOL, LTC. Tether (USDT) is accepted on Tron (TRC-20) and Ethereum (ERC-20); USD Coin (USDC) on Ethereum. There is no card, PayPal or bank payment. The crypto payments page lists the networks and the confirmations each one needs before your balance is credited.

You choose a top-up amount of at least $25, send the exact amount on the network shown, then click "I've sent the payment". The payment is matched by its exact amount and credited once it is confirmed on the blockchain. You then buy a credit pack from your balance.

Yes, as a credit to your balance. Within 72 hours of purchase, a credit pack is credited back to your balance if you used less than 20% of its searches. A pack that does not work, and that support cannot fix within 24 hours, is credited back in full. Top-ups are credited to your balance and are not paid back, except where the law requires it; the part added only to reach the minimum top-up is refundable on request within 30 days, in USDT or USDC, if unspent. The refund policy works through three examples.

Your API key appears in your dashboard as soon as you buy a credit pack from your confirmed crypto balance. API keys you can create and revoke at any time from your dashboard. Send the key in the Authorization: Bearer header, or as api_key for a quick test. Keep it in an environment variable such as SERPCRYPTO_API_KEY, never in code you share.

Request metadata for 30 days. We log, for each request, the time, account, engine, query parameters, status and credits used, and keep these logs for 30 days to show your usage and settle billing disputes, then delete them. We do not sell or share your queries. We keep your account number, orders, top-up records and credit totals while your account is open, and afterwards only as long as accounting rules require. We disclose records to authorities only when the law requires it. The privacy policy has the full table of what is kept and for how long.

No sale and no sharing for marketing or datasets. To run a search, our policy is to send upstream only the search parameters, never your account number, email or API key. Records are disclosed to authorities only when the law requires it, and logs older than 30 days no longer exist.

Your account number is your login: save it somewhere safe. Add an email (optional) if you want a way to recover it. Without an email, a lost account number cannot be recovered. If you added an email, recovery goes through it. Write the account number in a password manager the day you open the account.

Yes. The docs have copy-paste code for Python (requests and async httpx), Node.js (native fetch) and curl. There is no SDK to install: the API is one GET request with query parameters, which any HTTP client can send.

Yes. The MCP guide has a short Python server that exposes a web_search tool to Claude Desktop, Cursor and other MCP clients. The server calls https://serpcrypto.com/api/v1/search with your key from an environment variable. It is example code you run yourself, not a published package.

Yes. The LangChain page wraps the API as a tool for agents, and the OpenAI page gives the JSON schema of a web_search function and the loop that runs it. The same schema works with other tool-calling models.

Each account can run a set number of searches at the same time; the limit is shown in your dashboard and raised on request. A request above the limit gets HTTP 429 concurrency_limit and is not charged. GET https://serpcrypto.com/api/v1/credits returns your current limit. The Python and Node.js guides show how to cap parallel requests.

We do not quote a response time yet, because we have not measured one with a published method. Time a few requests yourself with the smallest pack, set a generous client timeout, and retry a 504 search_timeout after a short wait: it is not charged. A measured figure will appear on the status page with its method and date.

SerpGoat LLC runs SerpCrypto and sells its credits. Support answers at [email protected]: Email replies within 12 hours, 7 days a week, including help with API integration. Abuse reports go to [email protected]. The about page explains how results are collected and which figures we do not publish yet.

It depends on what you collect, how you use it and where you are, and some questions are being argued in US courts in 2026. Using search results for SEO, research and AI grounding is common practice. This is not legal advice: read our overview and ask a lawyer for your case.

Searches meant to find, track or harass a private person, child sexual abuse material, spam, fraud, and any use that breaks the law where you or we operate. Reselling raw API access as your own API needs a written agreement. The Acceptable Use Policy gives the full list. Abuse reports are reviewed by a person, and confirmed abuse leads to suspension.