Bing Search API Replacement — Bing Search Scraper

We do the dirty work so your dataset stays clean. 😈

$1.05 / 1,000 SERP rows — the drop-in replacement for the retired Bing Search API. Microsoft shut down the official Bing Search API on August 11, 2025, stranding SEO agencies, brand-monitoring tools, AI search agents, and RAG pipelines. This Actor is the successor: residential-proxy-backed, locale-aware, browser-fingerprint-rotating, and priced for high-volume workloads at a fraction of SerpApi or DataForSEO.

🎯 What this scrapes

Bing's public www.bing.com/search SERP — the same page any browser renders for a typed query. Per query the Actor:

1. Issues GET https://www.bing.com/search?q=<query>&cc=<country>&setlang=<language>&count=10&first=<offset> via curl-cffi with rotating browser fingerprints.
2. Parses every <li class="b_algo"> block with parsel to extract title, snippet, displayed URL, and the redirect-wrapped href.
3. Decodes the Bing redirect: bing.com/ck/a?...&u=a1<base64url> → canonical destination URL. No Bing redirect wrappers ever reach your dataset.
4. Paginates via &first=0,10,20,... until maxResultsPerQuery rows are emitted or the page returns zero organic blocks.

Each output row carries:

Field	Type	Description
query	string	The input search query that produced this row
position	integer	1-indexed organic rank within the query (across pages)
title	string	Result headline text
url	string	Canonical destination URL (decoded from bing.com/ck/a)
displayed_url	string | null	Breadcrumb-style URL as Bing renders it
snippet	string | null	Description excerpt shown under the result
country	string	ISO-3166-1 alpha-2 country passed in (cc=)
language	string	ISO-639-1 language passed in (setlang=)
scraped_at	string	ISO 8601 UTC timestamp of row creation

🔥 Features

• Drop-in Bing Search API replacement — same shape (query → list of organic results), no Microsoft API key required.
• Locale-aware — cc= / setlang= flow straight through to Bing; one Actor covers every market from US to JP to DE.
• URL decoded — Bing redirect wrappers (bing.com/ck/a?...) are always decoded to the canonical destination URL. Your dataset joins cleanly against any URL-keyed dataset.
• We rotate browser fingerprints — curl-cffi impersonates Chrome 131, Chrome 124, and Firefox 147 TLS+H2 fingerprints per page. Bing sees real-browser handshakes, not Python.
• We rotate residential proxies — Apify Proxy is on by default. Bing weights residential exits strongly over datacenter IPs, especially for paginated bursts. On every block we rotate session_id and exit IP.
• We retry with exponential backoff — 429 / 503 responses trigger a 2 s → 4 s → 8 s → 16 s → 30 s retry ladder, honouring Retry-After headers, up to 5 attempts per page.
• Per-query isolation — one query failing (zero results, network hiccup) does not abort the run. Remaining queries continue and the run surfaces a partial-success status message.
• Early-stop on empty page — we stop paginating when Bing stops returning organic blocks, instead of hammering empty pages.
• Pydantic v2 validation — input and every output row are model-validated. Invalid input fails fast with a clear message before any network call.

💡 Use cases

• Bing Search API migration — Microsoft retired the official endpoint on August 11, 2025. This Actor is the drop-in replacement for client reports, AI search agents, and competitive intelligence pipelines that depended on it.
• SEO rank tracking — schedule weekly runs across your target queries, write to a named dataset, and chart position over time per keyword and locale.
• Brand monitoring — watch the top 30 organic results for branded queries (your brand, competitors, key product names) and alert on new entrants.
• AI search agents and RAG pipelines — feed live SERP context into LLM agents. Bing diversifies your retrieval from Google: more forums, more long-tail, more long-form blogs.
• Multi-locale SERP comparison — pass the same query with different cc= / setlang= values to compare what Bing shows in the US vs UK vs DE vs JP.
• Content-gap analysis — pull the top 30 results for queries you want to rank for and reverse-engineer what content Bing rewards.
• Quality-assurance for Bing Webmaster Tools data — cross-check the impressions and positions Bing reports against what the live SERP actually shows.
• Bing SERP data for LangChain and LlamaIndex agents — replaces the now-dead BingSearchAPIWrapper and BingToolSpec integrations that called the retired API.

⚙️ How to use it

1. Open the Actor input form.
2. Enter at least one query in Queries (one row per query × organic result).
3. (Optional) Set Country — ISO-3166-1 alpha-2 (default US; common: GB, DE, FR, JP, BR).
4. (Optional) Set Language — ISO-639-1 (default en).
5. (Optional) Set Max results per query — 1 to 100 (default 30; pagination quality degrades past page 5–7).
6. Leave Use Apify Proxy on. Bing weights residential exits strongly over datacenter IPs.
7. Click Start. Results stream into the default dataset.

Quick examples

One query, first page only (the QA fixture):

{
  "queries": ["devilscrapes apify"],
  "country": "US",
  "language": "en",
  "maxResultsPerQuery": 10,
  "useProxy": true
}

Multi-query rank tracking (US, top 30 each):

{
  "queries": [
    "best web scraping tool",
    "apify alternatives",
    "bing search api replacement"
  ],
  "country": "US",
  "language": "en",
  "maxResultsPerQuery": 30,
  "useProxy": true
}

Multi-locale brand monitoring (top 20, three markets):

{
  "queries": ["devilscrapes"],
  "country": "GB",
  "language": "en",
  "maxResultsPerQuery": 20,
  "useProxy": true
}

Run three times, swapping country / language, and merge datasets on your side.

📥 Input

Field	Type	Required	Default	Description
queries	array of strings	yes	—	Search queries to run. One dataset row per query × organic result. At least one required.
country	string	no	US	ISO-3166-1 alpha-2 country passed to Bing as cc=.
language	string	no	en	ISO-639-1 language passed to Bing as setlang=.
maxResultsPerQuery	integer 1..100	no	30	Upper bound on organic results per query. Paginated in steps of 10.
useProxy	boolean	no	true	Route via Apify Proxy (residential group). Strongly recommended.

📤 Output

One dataset row per query × organic result. Example record:

{
  "query": "bing search api replacement",
  "position": 1,
  "title": "DevilScrapes — Bing Search Scraper",
  "url": "https://apify.com/DevilScrapes/bing-search-scraper",
  "displayed_url": "https://apify.com › DevilScrapes › bing-search-scraper",
  "snippet": "Drop-in replacement for the retired Bing Search API. Organic results for any query, locale-aware, residential-proxy-backed.",
  "country": "US",
  "language": "en",
  "scraped_at": "2026-05-16T13:40:00.000Z"
}

Download as JSON, CSV, Excel, or XML from the Export button on the run page.

💰 Pricing

Pay-Per-Event (PPE):

Event	Rate (USD)	Trigger
actor-start	$0.05	Once per Actor run at boot
result-row	$0.001	Per organic SERP row pushed

A typical run (10 queries × 30 results = 300 rows) costs ~$0.35. Per-1,000-row extrapolation: $1.05. Priced below typical Google SERP scrapers to capture teams migrating from the retired Bing Search API. You pay only for results that land — no data, no charge beyond the $0.05 start fee.

🚧 Limitations

• Organic web results only — People-Also-Ask boxes, Related Searches, news, images, videos, maps, and Bing's local pack are out of scope for v1.
• Featured snippets are not extracted as a separate field — when Bing's featured snippet sits inside a <li class="b_algo"> block it appears as a regular row. No is_featured flag in v1.
• Up to 100 results per query — maxResultsPerQuery is capped at 100. Past page 10 Bing's organic results become unreliable.
• No JavaScript rendering — we hit server-rendered HTML directly. If Bing migrates to a fully client-rendered SERP, a Camoufox path would be needed.
• Country / language are passed to Bing, not validated — cc= / setlang= are interpolated raw into the URL. An unrecognized value will silently return default-locale results. Use standard ISO codes.
• No historical tracking built in — every run is a fresh snapshot. Schedule runs and export to your own storage (BigQuery, S3, named Apify dataset) to build a time series. Apify default run-scoped storage is purged after 7 days on the free plan.

❓ FAQ

Why does this exist? Isn't there an official Bing Search API? Microsoft retired the Bing Search API on August 11, 2025. There is no longer an official programmatic endpoint for Bing organic results. The replacement Microsoft offers — "Grounding with Bing Search" inside Azure AI — is locked to a single LLM workflow and costs orders of magnitude more per call. This Actor hits the public SERP that any browser hits.

What does "bing search api replacement" mean in practice? The retired API returned a JSON payload of organic results for a query. This Actor returns the same shape: query in, list of {position, title, url, snippet, ...} rows out. Migration means swapping an API call for an Apify Actor run. The output schema is stable across versions.

Is this allowed? The Actor sends well-formed HTTP requests to a public, unauthenticated SERP endpoint and parses the publicly visible HTML. It respects HTTP retry semantics (429 / 503 / Retry-After). Your obligation as the user is to comply with Bing's Terms of Service in your jurisdiction. This Actor does not bypass any login, paywall, or authentication mechanism.

Why does the Actor rotate browser fingerprints? curl-cffi impersonates real browser TLS+H2 fingerprints. Rotating across Chrome 131, Chrome 124, and Firefox 147 per page reduces correlation between consecutive requests — which lowers the chance Bing flags a paginated burst as automated traffic. We handle the blocks so you don't have to.

Why is position 1-indexed? That is how every SEO tool, rank tracker, and BI dashboard counts. Position 1 = the first organic result. It drops straight into spreadsheets and databases without an off-by-one fix.

Why are URLs decoded? Can I get the raw bing.com/ck/a redirect? v1 always decodes. The redirect wrapper is a click-tracking detail; it adds no information your downstream workflow can use. Canonical URLs join cleanly with any other dataset keyed on URL.

Does this work for non-English locales? Yes. setlang=de / setlang=ja / setlang=fr all flow through to Bing. The parser is structural (li.b_algo blocks); it does not depend on English text patterns.

What about People-Also-Ask (PAA) boxes? Out of scope for v1. If there is demand, v2 will add an opt-in extractPaa: true input field that adds a separate kind: "paa" row per question.

How do I use this as a drop-in for LangChain's BingSearchAPIWrapper? Run the Actor with your query, export the dataset, and pass the url + snippet fields into your chain's retrieval step. The JSON output shape is close to what the dead API returned. A full cookbook migration example is in progress for Dev.to.

Do I need the Apify Proxy? Strongly recommended. Bing favours residential IPs over datacenter IPs, especially for paginated bursts. The default is ON. Turn it off only if you have your own proxy stack.

💬 Your feedback

Found a parser that broke after a Bing SERP markup change? Want a new feature (PAA boxes, news vertical, location targeting)? Open an issue on the Actor's Apify Store page or reach us at apify.com/DevilScrapes. We monitor publish-day QA failures and ship patches the same week.