Bing · 9.89%
Almost one US search in ten (StatCounter, September 2026). Worth tracking for US clients.
Proxies for SERP and rank tracking, the way local searchers see it
Rotate residential IPs in the searcher’s country and keep gl, hl and the location parameter consistent with the exit. From $2.70 per GB, plus ISP and datacenter IPs with unlimited traffic for crawling your own sites.
Choose your route
Proxies or a SERP API?
We do not sell a SERP API, so here is the honest version: both routes are legitimate, and they suit different teams.
SERP API
Someone else runs the scraper
- Rendering, parsing and blocks are handled for you; you get JSON back.
- Billed per 1,000 results, roughly $0.25–1.50 at large providers in 2026.
- A fixed menu of engines, result types and locations.
- Best if you do not run a tracker yet and want parsed data quickly.
A good choice if you would rather not run infrastructure.
ChronoProxy
Your tracker plus our proxies
- You already run a rank tracker, a scraper or a headless browser.
- You need engines, verticals or markets that an API does not cover.
- You want control over every request: device, parameters, pacing.
- You pay per GB of traffic, not per result: $4.80 → $2.70.
Proxies give you the network; your stack does the rest.
Dated changes
What changed for rank tracking in 2025–2026
Most proxy guides still describe the old Google. These are the changes that set today’s cost and method.
JavaScript becomes required
Google began requiring JavaScript to show results, citing protection against bots. Plain HTTP fetchers stopped working and headless browsers became the norm, at several times the compute.
Bing Search APIs retired
Microsoft retired the Bing Search APIs, removing a cheap official source of web results.
&num=100 stops working
Around 10–14 September Google stopped honouring the 100-results parameter and later said it was not something it formally supports. In a study of 319 properties, 87.7% saw Search Console impressions drop. A top-100 check now takes about ten requests per keyword, and many trackers cut depth to the top 20–50.
Google sues SerpApi
Google filed a DMCA suit against SerpApi over getting around its bot protection with fake browsers and many IP addresses.
AI Mode passes a billion users
Google reported AI Mode in 200+ countries with more than a billion monthly users. AI answers are now part of the results page to monitor.
Two rulings, different outcomes
A California court dismissed most of Google’s claims against SerpApi; eleven days later a New York court, in Reddit’s suit against SerpApi and Perplexity, held that SearchGuard is an access control under the DMCA. Promising to “bypass” Google is a legal risk, not a feature. The cases explained.
Custom Search JSON API ends
The API is already closed to new customers; existing users must migrate before the end date.
Localisation
How Google decides which results to show
Google says it estimates location from your IP, “roughly based on geography”, alongside device location and your history, always over an area of at least 3 km² shared by at least 1,000 users. For tracking, every signal you control should point to the same place.
| Signal | What it changes | What to set |
|---|---|---|
| IP address | Approximate location of the searcher | An exit in the target country, e.g. -cc-de |
gl parameter | The country results are for | The exit country |
hl parameter | Interface and query language | The market’s language: hl=de for Germany |
| Country domain | Less and less: Google announced in April 2025 that google.co.uk, google.de and others redirect to google.com | google.com with gl and hl |
uule parameter | Location down to a city (undocumented, reverse-engineered) | The city you track, inside the exit country |
| Device | Desktop vs mobile layout and ranking | A matching user agent and viewport |
| Signed-in history | Personalisation | Always track signed out, in a clean profile |
| Distance to the business | Local pack and map results | Pin the city with uule |
Our gateway targets by country today; region and city keys are planned for a later release. For city-level results, pair a country exit with uule.
Pick by task
Proxy type by SEO task
| Task | Proxy | Session | Notes |
|---|---|---|---|
| Daily Google rank tracking | Residential | Rotating | A new IP per request, or a short sticky session for a multi-page check. |
| Local and map-pack tracking | Residential | Rotating | Country exit plus uule for the city. |
| Mobile rankings | Residential or Mobile | Rotating or sticky | A mobile user agent on either; carrier exits cost the same, so compare both. |
| Fixed-location audits and long sessions | ISP | Static all month | The same address for repeatable checks. |
| Crawling your own sites, backlink and technical audits | Datacenter | Static | Unlimited traffic and up to 10 Gbps, but not for Google results. |
| Bing and other engines | Residential | Rotating | An exit in the engine’s main market. |
Devices
Desktop or mobile results?
Track the device your audience uses. In September 2026 StatCounter put mobile at 49.1% of US page views against 48.5% on desktop, at 53.6% in the UK and at 59.9% in Germany. For German or British clients, mobile rankings are the majority view.
A mobile user agent through a residential exit is the usual way to fetch mobile results. Since our mobile carrier exits cost the same per GB, it is cheap to test whether they change your results for the markets you track.
Beyond Google
Engines worth tracking
Yandex · ~70%
The main engine in Russia. Use an exit in the engine’s market for representative results.
Naver · ~43%
Close to Google’s share in Korea. Local results differ sharply between the two.
Ecosia · ~1%
Small but visible in Germany, next to Yandex at 1.56% and Bing at 6.49%.
Pacing
Rotation and pacing rules that keep data clean
- One request per IP for Google, or a short sticky session for pages of the same query.
- Keep exit country, gl, hl and uule consistent. A mismatch gives you a page no real user sees.
- Back off on a CAPTCHA or a 429. Pause and slow down; do not solve CAPTCHAs at scale.
- Track signed out in a clean profile with no history.
- Cache results and do not re-check a keyword more often than the decision needs.
- Log the exit IP and its geolocation next to every result.
- Budget for depth: a top-100 check now costs about ten requests.
- No fake clicks and no ranking manipulation. Observe; do not interfere.
For developers
Check the exit before you trust a ranking
Verify where the request leaves from, then hand your tracker matching parameters. The snippets stop there on purpose: no CAPTCHA handling, no protection workarounds.
import requests, urllib.parse
proxy = "http://cp_a1b2c3-cc-de-sid-serp01-ttl-10:[email protected]:7000"
geo = requests.get("https://ipinfo.io/json", proxies={"http": proxy, "https": proxy}, timeout=15).json()
print(geo["country"], geo.get("city"), geo.get("org")) # DE Cologne AS3320 Deutsche Telekom AG
# parameters for your tracker, so IP, country and language agree
params = {"q": "laufschuhe", "gl": "de", "hl": "de", "pws": "0"}
print("https://www.google.de/search?" + urllib.parse.urlencode(params))import { ProxyAgent } from "undici";
const login = "cp_a1b2c3-cc-gb-sid-serp02-ttl-10";
const dispatcher = new ProxyAgent({
uri: "http://gate.chronoproxy.com:7000",
token: "Basic " + Buffer.from(`${login}:PASS`).toString("base64"),
});
const geo = await (await fetch("https://ipinfo.io/json", { dispatcher })).json();
console.log(geo.country, geo.city, geo.org); // GB Manchester AS5089 Virgin Media Limitedcurl -x gate.chronoproxy.com:7000 \
-U "cp_a1b2c3-cc-us:PASS" \
https://ipinfo.io/jsonBudget
Requests, gigabytes and cost
Rank tracking is arithmetic: keywords × locations × devices × pages × how often. Page weight decides the gigabytes, so it is the number to measure on your own stack.
DevicesDepthHow oftenPage weight
Requests per month300,000
Traffic per month88 GB
Residential cost per month$342.77
$3.90 per GB at this volume.
For comparison, a SERP API at $0.25–1.50 per 1,000 results would cost about $75.00–$450.00 for the same requests.
Get residential proxiesAt full rendering a SERP API can be cheaper than proxies. Proxies win when you block images and scripts, need markets an API lacks, or already run your own tracker.
Pricing
What SERP work costs here
| Proxy | SEO job | Session | Price |
|---|---|---|---|
| Residential | Rank tracking and local results | Rotating or sticky | $4.80 → $2.70 per GB |
| Mobile | Mobile rankings | Rotating or sticky | $4.80 → $2.70 per GB |
| ISP | Repeatable fixed-location audits | Static all month | $2.00 per IP per month |
| Datacenter | Crawling your own sites | Static | $1.20 per IP per month |
Know what Google says about automated queries
Google’s spam policies count “scraping results for rank-checking purposes” as machine-generated traffic, and its terms restrict automated access. Rank tracking through proxies is common industry practice, but it is your responsibility: public results only, no fake clicks or ranking manipulation, and no CAPTCHA solving at scale.
Not legal advice. ChronoProxy is not affiliated with Google or Microsoft.SERP proxy questions
Proxies used by rank trackers and SEO tools to request search results from the location being tracked. Rotating residential IPs are the usual choice because they look like ordinary searchers.
Rotating residential for Google rankings and local results, mobile or residential with a mobile user agent for mobile rankings, ISP for repeatable audits and datacenter only for crawling your own sites.
An API if you want parsed results without running a scraper. Proxies if you already run a tracker, need markets or engines an API does not cover, or want control over each request.
Yes, by pairing a country exit with the uule location parameter for the city. Our gateway targets by country today, with region and city keys planned for a later release. ZIP-level exits are not offered.
For local packs it still helps: uule states the location explicitly, while IP location is approximate. Keep them consistent so they never point to different places.
Not necessarily. A mobile user agent on a residential exit returns mobile results; our carrier exits cost the same, so test both for your markets.
For Google, one request per IP, or a short sticky session when you fetch several pages of the same query.
Too many requests too quickly, or signals that do not add up. Slow down, check that IP, gl and hl agree, and back off rather than retrying at once.
Requests per month times page weight. 500 keywords × 5 locations × 2 devices × 2 pages daily is 300,000 requests a month, about 88 GB at 0.3 MB per page. The calculator does it for your numbers.
Since September 2025 a top-100 check takes about ten paginated requests instead of one, which multiplies traffic and is why many tools now track the top 20–50.
Any tool that accepts an HTTP or SOCKS5 proxy with a login and password can use the gateway at gate.chronoproxy.com, port 7000 or 7001.
Yes. Use an exit in the engine’s main market: the US for Bing, Russia for Yandex, South Korea for Naver.
Your tracker can record them if it renders the page with JavaScript; the proxy only supplies the location. Expect heavier pages when AI answers appear.
Google’s policies treat automated rank checking as machine-generated traffic and its terms restrict automated access. Many companies do it anyway; assess the risk for your use. This is not legal advice.