Skip to content

Bing Ads SERP Scraping API for Placement Data

Polling Bing search results for sponsored placements by hand runs into rate limits and CAPTCHAs, and the markup changes often enough that a home-grown scraper becomes a maintenance job.

OUTPUT FORMATSHTMLJSONMarkdown

This page describes a hosted API that renders the Bing SERP, rotates proxies and returns either rendered HTML or structured JSON, so the work shifts from infrastructure to defining what you want to extract. It is aimed at growth teams tracking competitor creatives, analysts measuring share of voice, and developers wiring ad monitoring into an existing stack.

What the API actually returns

You send a Bing search-results URL that carries sponsored placements. The service loads the page in a headless browser, executes its JavaScript, and returns the rendered HTML by default. If you supply extraction rules, it returns structured JSON instead. The public surface of a sponsored result is consistent enough to plan around: ad headline, display URL, call-out extensions, sitelinks, slot position and the landing URL. Those fields are what a competitor ad library or a share-of-voice calculation is built from.

The distinction worth holding on to: this is a way to read the public Bing SERP. It is not the Microsoft Advertising reporting API, and it does not give you advertiser-side metrics such as spend, click-through rate or conversion data. If you need those, you need the official advertising platform, not a scraper.

Three jobs this fits

Most teams arrive with one of three goals, and each maps to a different extraction rule set.

Tracking sponsored placements. You keep a list of keywords and want to know which advertisers appear, in which slot, and how that changes over time. The useful output is a per-keyword record of advertiser, position and timestamp.

Comparing ad creatives. You want the headline copy, display URL, call-outs and sitelinks for each sponsored result, plus how often the creative changes. That becomes a competitor ad library you can review on a schedule.

Monitoring share of voice. You aggregate appearances across many queries to see which advertisers dominate sponsored slots in a category. This is the most volume-hungry job, because a single query tells you almost nothing; you need breadth and repeated sampling.

Field map for a sponsored placement

Before writing extraction rules, decide which fields you actually need. A practical minimum for competitive tracking is headline, display URL and slot position; add call-outs and sitelinks if creative comparison is the goal, and add the landing URL if you plan to follow the click path. Creative refresh frequency is not a field on the page — it is derived by comparing snapshots over time, so your storage needs a timestamp on every record.

A common mistake is to write one selector set and assume it holds across every keyword. Sponsored blocks vary in how many results they contain and whether extensions render at all. Build the rule set so a missing extension produces an empty value rather than a failed parse.

Worked example: one keyword, one request

Take a single tracked keyword. You build the Bing search URL for it, send it to the API, and ask for JSON back using extraction rules that target the sponsored block. The response gives you the advertisers present, their slot order, and their creative text for that moment.

The cost of that request depends on which options you enable, not on the keyword. A plain request is the cheapest. Enabling JavaScript rendering raises the cost. Routing through premium proxies raises it further, and the premium-proxy cost differs depending on whether JavaScript rendering is also on. Stealth proxy is the most expensive option. Because Bing SERPs are JavaScript-heavy and protected, most real tracking runs will need at least rendering, so budget from the rendered cost rather than the base rate.

To turn one request into a share-of-voice number, repeat it across your keyword list on a fixed schedule and store each snapshot with a timestamp. The metric is then a count of appearances per advertiser divided by total sponsored slots observed in the period.

Request options and what they change

The API exposes a small set of parameters that matter for this job. render_js turns on JavaScript execution, which Bing SERPs generally need. stealth_proxy is the option aimed at CAPTCHA-heavy targets. country_code sets the geographic origin of the request, which changes which ads you see. extract_rules takes CSS or XPath selectors and returns structured data; ai_extract_rules lets you describe the fields in natural language instead. Screenshots and JavaScript scenarios are available when you need to see the page as rendered or interact with it before capture.

A useful control map: if your output is wrong but the request succeeds, the problem is almost always the extraction rules, not the anti-bot layer. If the request fails or returns a challenge page, the problem is on the access side, and the lever is the proxy or rendering option rather than the selectors. Separating those two failure modes early saves a lot of debugging time.

Plans, credits and concurrency

The published plan ladder runs from a small hobby tier up through enterprise tiers, with API credits and concurrent requests rising at each step. JavaScript rendering, rotating and premium proxies, geotargeting, and screenshots with extraction rules are included from the entry plan. Priority email support, a dedicated account manager and team management appear only at the higher tiers. Credits are valid for one month and do not roll over, and requests start at one credit each before option costs are added.

The plan figures below are the captured reference values and should be treated as a dated snapshot; check the current pricing page before committing to a tier. All prices are shown exclusive of VAT.

Getting from zero to a first request

Create an account to get an API key and the free trial credits. Install the SDK for your stack — Python, Node.js, Go, PHP, Ruby and Java are published, and plain cURL works too — or skip the SDK and send an HTTP request directly. Point the request at a Bing SERP URL with sponsored placements. Start with rendered HTML so you can see the page structure, write your extraction rules against it, then switch to structured JSON for production runs.

A sensible first milestone is not 'scrape everything'. It is one keyword, one request, one saved JSON record with the fields you care about. Once that is stable, widen the keyword list and add scheduling.

Plan reference (captured snapshot)
PlanPrice / monthAPI creditsConcurrent requestsPriority supportAccount managerTeam management
Hobby$1975,00025Not includedNot includedNot included
Freelance$49250,00050Not includedNot includedNot included
Startup$991,000,000100Not includedNot includedNot included
Business$2493,000,000200IncludedIncludedIncluded
Business +$5998,000,000400IncludedIncludedIncluded
Enterprise 14$99914,000,000500IncludedIncludedIncluded
Enterprise 24$1,59924,000,000600IncludedIncludedIncluded
Enterprise 41$2,39941,000,000700IncludedIncludedIncluded
Enterprise 69$3,59969,000,000800IncludedIncludedIncluded
Enterprise 120$5,799120,000,000900IncludedIncludedIncluded

Captured reference values. Prices exclude VAT. Credits are valid for one month and do not roll over. JavaScript rendering, rotating and premium proxies, geotargeting, and screenshots with extraction rules are included on every plan listed. Check the current pricing page before choosing a tier.

Credit cost per successful request by option
Option combinationCredits per request
Base request1
JavaScript rendering enabled5
Premium proxies, JS rendering disabled10
Premium proxies with JS rendering25
Stealth proxy75

Costs apply per successful request and are additive to the base rate where the provider describes them as such. Verify against current documentation before sizing a budget.

Frequently asked questions

Do I need a Microsoft Advertising account to use this?

No. You need an account with the scraping provider to get an API key and the free trial credits. The service reads public Bing search results, so no advertiser-side login is involved.

Is collecting Bing sponsored placements legal?

Public-facing pages can usually be collected, but the target site's terms of service govern what you may do with the data. Review those terms before deploying, and avoid anything behind a login. The provider also states that it does not store personal data from scraped sites unless you explicitly request it.

What data can I get back?

Anything that renders on the public page. In practice that means ad headline, display URL, call-out extensions, sitelinks, slot position and landing URL. You get structured JSON by supplying extract_rules with CSS or XPath selectors, or ai_extract_rules for a natural-language description of the fields.

How do I make a first request?

Create an account, copy the API key, and send a GET request with the Bing SERP URL. The service renders the page, rotates proxies and returns the result. Start with rendered HTML to inspect the structure, then move to extraction rules once your selectors are stable.

Which languages and SDKs are supported?

Published SDKs cover Python, Node.js, Go, PHP, Ruby and Java. Plain cURL works for any other stack, so the choice is really about what your codebase already uses.

How is a request priced?

Every request starts at one credit, and options add to that. JavaScript rendering costs five credits. Premium proxies cost ten credits without rendering and twenty-five with it. Stealth proxy costs seventy-five. Credits are valid for one month and do not roll over, and paid plans start at $19 per month.

How is this different from the Microsoft Advertising API?

The official advertising API gives you advertiser-side reporting for accounts you control. This service reads the public Bing results page, so it can show you competitors' sponsored placements but not their spend, click-through rate or conversion data. The two answer different questions and are not substitutes.

My request succeeds but the JSON is empty. What is wrong?

That is almost always an extraction-rule problem rather than a blocking problem. The page loaded, but your selectors did not match the sponsored block for that keyword, or the block rendered without the extension you targeted. A failed or challenge response is the opposite case: the access layer needs a different proxy or rendering option.