Bing Knowledge Graph Scraper API
Our Bing knowledge graph scraper reads the answer box and the right-rail entity panel Bing shows for a query, returning the answer, secondary value, entity title, attributes, image, and official website as JSON on the same call that fetches the organic results.
Why Bing makes Bing Knowledge Graph data hard
The answer box and entity panel are where Bing puts the direct facts, but there is no API for them and Bing gates most of the entity card behind client-side hydration. The fact-style answer box server-renders on measurement and quick-fact queries; the full knowledge panel renders only when Bing chooses to. We parse whichever is present and return it as clean JSON, never failing the call over a missing block.
The Bing Knowledge Graph Scraper API, one request away
curl "https://api.bingscraperapi.com/api/v1/bing/search?q=how+tall+is+mount+everest&api_key=$API_KEY"import requests
BASE = "https://api.bingscraperapi.com"
API_KEY = "YOUR_API_KEY"
data = requests.get(
f"{BASE}/api/v1/bing/search",
params={"q": "how tall is mount everest", "api_key": API_KEY},
timeout=30,
).json()
# The featured answer / fact box (present on quick-fact queries).
box = data.get("answer_box")
if box:
print("answer:", box["answer"])
# The right-rail entity panel (present when Bing renders it).
kg = data.get("knowledge_graph")
if kg:
print(kg["title"], "-", kg.get("type"))
for k, v in kg.get("attributes", {}).items():
print(" ", k, ":", v)Parameters
| Parameter | Required | Default | Notes |
|---|---|---|---|
q | required | - | The query or entity name, e.g. how tall is mount everest or mount everest. This drives which answer box and panel Bing shows. |
url | optional | - | A full bing.com/search URL to parse instead of building one from q. Pass q or url. |
country | optional | US | Two-letter market region, e.g. GB or DE. Sets mkt=en-XX so the panel is the stable national one. Defaults to US. |
count | optional | 10 | Number of organic results to request alongside the blocks, 1 to 50. Defaults to 10. |
api_key | required | - | Your API key, passed as a query parameter. Get one free at signup. |
Fields returned by the Bing Knowledge Graph Scraper API
{
"query": "how tall is mount everest",
"engine": "bing",
"results_count": 8,
"answer_box": {
"type": "answer",
"answer": "8,849 m / 29,032 ft",
"title": "Deep dive into how tall is mount eve…"
},
"knowledge_graph": {
"title": "Mount Everest",
"type": "Mountain",
"description": "Earth's highest mountain above sea level, on the China-Nepal border.",
"attributes": {
"Elevation": "8,849 m",
"First ascent": "May 29, 1953",
"Mountain range": "Himalayas, Mahalangur Himal"
},
"image": "https://www.bing.com/th?id=...",
"website": "https://whc.unesco.org/en/list/120",
"source": "bing"
}
}| Field | Type | Description |
|---|---|---|
answer_box | object | The featured answer or fact box. Present on quick-fact queries; omitted otherwise. |
answer_box.answer | string | The direct answer text, e.g. a fact-table value like 8,849 m / 29,032 ft. |
answer_box.secondary | string | A secondary focus value when Bing shows one alongside the main answer. |
answer_box.title | string | The label or heading Bing prints above the answer. |
knowledge_graph | object | The right-rail entity panel. Present only when Bing server-renders it for the query. |
knowledge_graph.title | string | The entity name, e.g. Mount Everest. |
knowledge_graph.type | string | The entity type or subtitle, e.g. Mountain. |
knowledge_graph.attributes | object | Label-value pairs from the panel's fact table, e.g. Elevation, First ascent. |
knowledge_graph.website | string | The entity's official website, decoded from Bing's redirect, when shown. |
What this data is good for
Answer and fact capture
Entity enrichment
Featured-snippet monitoring
Knowledge base building
Brand and entity checks
Assistant grounding
Why our Bing Knowledge Graph Scraper API is reliable
One call returns the organic results plus the answer box and, when Bing renders it, the entity knowledge panel, each parsed into clean JSON with the official website decoded from Bing's redirect. Every parser is defensive, so a missing block returns nothing rather than erroring, and the request runs through rotating proxies with anti-bot handling and retries at a 2.6s median.
Answer box parsed
Entity panel when present
Blocks on the organic call
Decoded website link
Defensive, never fails on a block
Anti-bot and proxy rotation
Choosing between the Bing Knowledge Graph Scraper API and the Bing API
| Our API | DIY (requests / headless) | Bing Web Search API (Azure) | |
|---|---|---|---|
| Answer box | Parsed on fact queries | You write the selector | Some answers via computation |
| Entity panel | Returned when Bing renders it | You parse the entity card | Entities via a separate API |
| Availability | Live and maintained | Depends on your setup | Closed to new customers, sunsetting |
| Website link | Decoded to the real URL | You decode ck/a redirects yourself | Clean URLs |
| Setup | API key only | Proxies, headless browser, parsers | Azure account and resource keys |
| Delivery | On the organic call, no extra request | Extra parsing per block | Separate response fields |
Straightforward pricing
| Plan | Price | Best for |
|---|---|---|
| Free | 1,000 requests | Testing and small jobs |
| Pro | $0.60 / 1k | Production workloads |
| Pay-as-you-go | $0.90 / 1k | Spiky or one-off volume |
Median response 2.6s. You only pay for successful requests.
FAQ
A Bing knowledge graph scraper reads the answer box and the right-rail entity panel Bing shows for a query and returns them in a structured format. Our API returns the answer box with its direct answer on quick-fact queries, and when Bing renders the entity panel it returns the title, type, description, attribute table, image, and official website as JSON, on the same call as the organic results.
No. Bing gates most of the entity card behind client-side hydration and server-renders the full panel only when it chooses to, so the knowledge_graph object is present on some queries and omitted on others. The fact-style answer box renders more reliably on measurement and quick-fact queries. Our parser returns whichever block Bing serves and never fails the request over a missing one.
The answer box is the direct answer Bing prints for a factual query, for example a height or a date, returned in the answer_box object. The knowledge graph is the right-rail entity panel with a title, type, description, and an attribute table, returned in the knowledge_graph object. A query can trigger one, both, or neither, and our endpoint returns whatever Bing renders.
Measurement and quick-fact queries are the most reliable, for example how tall, how far, when did, or a unit conversion. Bing server-renders these as a fact table, which our parser reads into answer_box. Broader or ambiguous queries may return no answer box, in which case the field is simply omitted and you fall back to the organic results.
Yes. When the panel includes an official website, our scraper decodes it from Bing's ck/a redirect wrapper and returns the real destination URL in knowledge_graph.website, so it points straight at the entity's site rather than a bing.com tracking link.
Call the bing/search endpoint with your query. The same response carries the organic results plus the answer_box and, when present, the knowledge_graph object. You read the fields you need. There is no separate endpoint or extra request for the panel, and each call counts as one request against your plan.