API Reference — Scrape, SERP, Crawl

Quickstart

Get a key — register, no card. Keys start with qd_live_.
Call scrape — POST JSON to /v1/scrape. Copy, replace the key, run it.
Read payload — Markdown is in payload.content. payload.usage is what the call cost.

curl -X POST "https://api.quanticdata.io/v1/scrape" \
  -H "Authorization: Bearer qd_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url": "https://example.com", "format": "markdown"}'

import requests

r = requests.post(
    "https://api.quanticdata.io/v1/scrape",
    headers={"Authorization": "Bearer qd_live_YOUR_KEY"},
    json={"url": "https://example.com", "format": "markdown"},
    timeout=90,
)
data = r.json()["payload"]
print(data["title"], data["content"][:200])

const res = await fetch("https://api.quanticdata.io/v1/scrape", {
  method: "POST",
  headers: {
    Authorization: "Bearer qd_live_YOUR_KEY",
    "Content-Type": "application/json",
  },
  body: JSON.stringify({ url: "https://example.com", format: "markdown" }),
});
const { payload } = await res.json();
console.log(payload.title, payload.content.slice(0, 200));

The QuanticData API reference as a map: scrape, SERP, map, crawl, batch, SEO audit, collectors and proxies with their unit prices, all around one Bearer key and one host. — Eight endpoint groups, one host and one Bearer key: the price beside each endpoint is what a unit costs, and whichever one you call, the answer is the same envelope.

Authentication

Host https://api.quanticdata.io/v1. Every request sends Authorization: Bearer qd_live_… and Content-Type: application/json. One key covers every Data API and proxy generation. x-api-key is accepted as a fallback; do not put the key in the query string in production.

To import this API (Postman, Insomnia, Stoplight and similar): OpenAPI 3.0 at https://quanticdata.io/docs/openapi.yaml (JSON twin openapi.json) or the Postman Collection v2.1 at https://quanticdata.io/docs/postman.json. GraphQL and Kafka are not offered — REST only.

PAYG includes $2 of free usage every month and 60 requests per minute. Subscription tiers raise the cap and discount unit prices — see pricing and plans.

Envelope

Success is {"type":"response","message":"…","payload":{…}}. Failure is {"type":"error","message":"…"} and costs nothing. List endpoints also fill pagination. Billed calls add payload.usage (cost_usd, free_usd, paid_usd).

Response

{
  "type": "response",
  "message": "Extraction successful",
  "payload": { "content": "…", "engine": "tls", "usage": { "cost_usd": 0.0002 } }
}

Scrape

POST /v1/scrape alias /v1/scraper/extract $0.0002 · render $0.001

One URL through the residential pool, returned as Markdown, HTML or text. Starts on a TLS-fingerprint tier and escalates to a stealth browser only on a block, or when you set render: true.

Param	Type	Default	Description
url req	string	required	The page to fetch (http/https, max 2048 chars). Optional when you pass html instead.
format	string	markdown	markdown \| html \| text. format: "raw" is accepted as an alias for html.
formats	string[]	—	Extra formats to return together under formats in the response.
render	boolean	false	Force the stealth headless browser (JS execution). Bills the rendered rate.
engine	string	auto	auto \| tls \| fetch \| render. auto starts on the cheap TLS tier and escalates to a browser only on a block; tls never escalates.
mobile	boolean	false	Mobile viewport and user agent (render mode).
screenshot	bool \| "fullPage"	off	Base64 PNG of the page, render mode: true = viewport, "fullPage" = whole page.
country / state / city	string	—	Proxy exit geo (ISO country code; state/city require country).
rotation	string	rotating	rotating (new IP per request) \| sticky.
sessionId	string	auto	Sticky session id (max 64 chars, generated if omitted).
sessionDuration	number	10	Sticky session lifetime in minutes (3–1440).
contentMode	string	smart	smart (page minus nav/footer/cookie chrome) \| article (Readability main article) \| full (whole body). Alias: content_mode.
include_links	boolean	false	Also return the page's de-duplicated absolute links. Alias: includeLinks.
waitMs	number	—	Extra wait after load before capture (render mode, max 15000).
waitForSelector	string	—	Wait until this CSS selector appears (render mode).
scrollToBottom	boolean	false	Auto-scroll to trigger lazy-loaded content (render mode).
actions	object[]	—	Ordered interactions before capture (max 20), e.g. {"click": "#accept"}, {"scroll": "bottom"}.
xhr	boolean	false	Record the page's XHR/fetch traffic under xhr (forces render).
extract	object	—	CSS selector schema, e.g. {"price": ".price"} — returns a structured data object.
headers	object	—	Extra request headers replayed to the target (max 32; hop-by-hop headers rejected).
cookies	object	—	name → value cookies sent to the target (max 50).
mode	string	full	"summary" returns metadata only (title, description, canonical…) — the cheap mode for audits.
ai_prompt / ai_schema	string / object	—	LLM extraction: natural-language prompt and/or JSON Schema. Result under payload.ai.data.
html	string	—	Convert markup you already have (max 5MB) — no fetch, no proxy bandwidth.

Request

curl -X POST "https://api.quanticdata.io/v1/scrape" \
  -H "Authorization: Bearer qd_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url": "https://example.com", "format": "markdown"}'

import requests

r = requests.post(
    "https://api.quanticdata.io/v1/scrape",
    headers={"Authorization": "Bearer qd_live_YOUR_KEY"},
    json={"url": "https://example.com", "format": "markdown"},
    timeout=90,
)
data = r.json()["payload"]
print(data["title"], data["content"][:200])

const res = await fetch("https://api.quanticdata.io/v1/scrape", {
  method: "POST",
  headers: {
    Authorization: "Bearer qd_live_YOUR_KEY",
    "Content-Type": "application/json",
  },
  body: JSON.stringify({ url: "https://example.com", format: "markdown" }),
});
const { payload } = await res.json();
console.log(payload.title, payload.content.slice(0, 200));

Response

{
  "type": "response",
  "message": "Extraction successful",
  "payload": {
    "url": "https://example.com",
    "finalUrl": "https://example.com/",
    "status": 200,
    "title": "Example Domain",
    "format": "markdown",
    "content": "# Example Domain\n\nThis domain is for use in…",
    "engine": "tls",
    "usage": { "cost_usd": 0.0002, "free_usd": 0.0002, "paid_usd": 0 }
  }
}

Optional fields appear only when requested: data (extract), links (include_links), screenshot, xhr, formats, ai. Guide: Web Scraping API.

SERP

POST /v1/serp alias /v1/scraper/serp from $0.0005 · render $0.002

Structured search results: organic, ads, People Also Ask, related searches, knowledge panels and Google AI Overviews, across 17 verticals on Google, Bing and DuckDuckGo.

Param	Type	Default	Description
query req	string	required*	The search. *Optional for ID-addressed verticals: place_details (with place_id/data_id), product (product_id), flights, lens, reviews (data_id).
engine	string	google	google \| bing \| duckduckgo.
search_type	string	search	search, shopping, images, news, places, maps, videos, scholar, jobs, autocomplete, place_details, hotels, flights, events, product, lens, reviews, trends. Alias: type.
country	string	—	ISO country — proxy exit and engine locale. Alias: gl.
lang	string	—	UI language, e.g. "en", "it". Alias: hl.
num	number	10	Results to request, clamped to 1–100.
page	number	1	Result page, 1-based. start (offset) is also accepted.
device	string	desktop	desktop \| mobile.
render	boolean	per engine	Google renders by default; pass false for the cheaper HTTP tier. Bing/DuckDuckGo do not render by default.
location	string	—	Human-readable search location ("Milan, Italy") — encoded to uule server-side.
uule	string	—	Encoded uule token OR raw "lat,lon[,radius]" coordinates.
safe	string	—	active \| off — SafeSearch.
nfpr	boolean	—	1/true disables auto-corrected results.
google_params	object	—	Escape hatch: extra Google URL params (max 24 keys).

Vertical-specific params: place_id / data_id, product_id, flight dates and airport ids, hotel dates, image_url (lens), gps_coordinates (maps), next_page_token (reviews).

Request

curl -X POST "https://api.quanticdata.io/v1/serp" \
  -H "Authorization: Bearer qd_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"query": "best coffee grinder", "engine": "google", "country": "us"}'

import requests

r = requests.post(
    "https://api.quanticdata.io/v1/serp",
    headers={"Authorization": "Bearer qd_live_YOUR_KEY"},
    json={"query": "best coffee grinder", "engine": "google", "country": "us"},
    timeout=90,
)
for row in r.json()["payload"]["organic"]:
    print(row["rank"], row["title"], row["link"])

const res = await fetch("https://api.quanticdata.io/v1/serp", {
  method: "POST",
  headers: {
    Authorization: "Bearer qd_live_YOUR_KEY",
    "Content-Type": "application/json",
  },
  body: JSON.stringify({ query: "best coffee grinder", engine: "google", country: "us" }),
});
const { payload } = await res.json();
for (const row of payload.organic) console.log(row.rank, row.title, row.link);

Response

{
  "type": "response",
  "message": "SERP successful",
  "payload": {
    "search_parameters": { "engine": "google", "q": "best coffee grinder", "country": "us" },
    "organic": [
      { "rank": 1, "title": "…", "link": "https://…", "description": "…" }
    ],
    "people_also_ask": [ … ],
    "usage": { "cost_usd": 0.0005, "free_usd": 0.0005, "paid_usd": 0 }
  }
}

Guide: SERP API.

Map

POST /v1/map alias /v1/scraper/map $0.0005

URL discovery for a whole site: robots.txt sitemaps, /sitemap.xml (including nested indexes) and same-domain homepage links, de-duplicated. Synchronous — no per-page crawl.

Param	Type	Default	Description
url req	string	required	Seed URL — the site to map.
limit	number	100	Max URLs to return (cap 5000). The response always carries the site-wide total and a per-section summary.
search	string	—	Only return URLs containing this substring.
includeSubdomains	boolean	false	Include URLs on subdomains of the seed host.
sitemapOnly	boolean	false	Skip the homepage link scrape.
group_by	string	—	"path" returns the path tree with counts instead of the URL list.
country	string	—	ISO country code for the proxy exit geo.

Request

curl -X POST "https://api.quanticdata.io/v1/map" \
  -H "Authorization: Bearer qd_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url": "https://example.com", "limit": 100}'

Response

{
  "type": "response",
  "message": "Map successful",
  "payload": {
    "url": "https://example.com",
    "links": ["https://example.com/", "https://example.com/about"],
    "count": 100,
    "total": 2417,
    "summary": { "/blog": 1988, "/docs": 240 },
    "usage": { "cost_usd": 0.0005, "free_usd": 0.0005, "paid_usd": 0 }
  }
}

Guide: Crawl & Map.

Crawl

POST /v1/crawl alias /v1/scraper/crawl

GET /v1/crawl/{jobId} alias /v1/scraper/crawl/{jobId}

DELETE /v1/crawl/{jobId} alias /v1/scraper/crawl/{jobId} $0.0003 / page

Asynchronous BFS from a seed. POST returns a job id; poll GET. Use ?since={nextCursor}&include_content=false for cheap status checks, then fetch bodies when the job completes.

Param	Type	Default	Description
url req	string	required	Seed URL the crawl starts from.
limit	number	50	Max pages (cap 500). Charged up front on this limit; unfetched or failed pages are refunded when the job settles.
depth	number	3	Max link depth from the seed (cap 10).
format	string	markdown	markdown \| html \| text, applied to every page.
contentMode	string	smart	smart \| article \| full. Alias: content_mode.
render	boolean	false	Render each page with the stealth browser (rendered rate per page).
sameDomain	boolean	true	Only follow same-host links.
allowSubdomains	boolean	false	Also follow subdomains of the seed host.
include / exclude	string[]	—	URL substrings/globs to include or exclude (max 50 patterns).
country	string	—	ISO country code for the proxy exit geo.

Request

curl -X POST "https://api.quanticdata.io/v1/crawl" \
  -H "Authorization: Bearer qd_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url": "https://example.com", "limit": 20, "depth": 2}'

# poll, then cancel
curl "https://api.quanticdata.io/v1/crawl/JOB_ID?include_content=false" \
  -H "Authorization: Bearer qd_live_YOUR_KEY"
curl -X DELETE "https://api.quanticdata.io/v1/crawl/JOB_ID" \
  -H "Authorization: Bearer qd_live_YOUR_KEY"

Response

{
  "type": "response",
  "message": "Crawl started",
  "payload": {
    "id": "d5b8…", "status": "running", "seed": "https://example.com/",
    "limit": 20, "pagesCrawled": 0, "pagesQueued": 1,
    "statusUrl": "/api/v1/scraper/crawl/d5b8…"
  }
}

Guide: Crawl & Map.

Batch

POST /v1/batch alias /v1/scraper/batch

GET /v1/batch/{jobId} alias /v1/scraper/batch/{jobId} $0.0002 / URL

Scrape up to 5,000 URLs you already know, with shared options. POST returns a job id. Poll GET with include_content=true and since, or pass webhook.

Param	Type	Default	Description
urls req	string[]	required	URLs to scrape (max 5000, 2048 chars each). Charged up front per URL; the failed share is refunded at settlement.
format	string	markdown	markdown \| html \| text.
engine	string	auto	auto \| tls \| fetch \| render.
render	boolean	false	Force the headless browser for every URL (rendered rate per URL).
extract	object	—	CSS selector schema applied to every URL.
contentMode	string	smart	smart \| article \| full. Alias: content_mode.
mode	string	full	"summary" stores per-URL metadata only.
concurrency	number	5	Simultaneous fetches (cap 20).
webhook	string	—	Public http(s) URL to POST the finished job to.
country	string	—	ISO country code for the proxy exit geo.

Request

curl -X POST "https://api.quanticdata.io/v1/batch" \
  -H "Authorization: Bearer qd_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"urls": ["https://example.com", "https://example.org"], "format": "markdown"}'

curl "https://api.quanticdata.io/v1/batch/JOB_ID?include_content=true" \
  -H "Authorization: Bearer qd_live_YOUR_KEY"

SEO audit

POST /v1/seo-audit alias /v1/scraper/seo-audit $0.0012

Fetches a URL twice — pure HTTP (no JS) and fully rendered — and returns both views, the diff (JS-only content, changed title/description, missing canonical) and bot-facing meta (robots, OG, JSON-LD types).

Param	Type	Default	Description
url req	string	required	The page to audit.
no_render	boolean	false	Skip the rendered pass — no-JS view only, bills at the plain scrape rate. Alias: noRender.
country	string	—	ISO country code for the proxy exit geo.

Request

curl -X POST "https://api.quanticdata.io/v1/seo-audit" \
  -H "Authorization: Bearer qd_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url": "https://example.com"}'

Response

{
  "type": "response",
  "message": "SEO audit successful",
  "payload": {
    "url": "https://example.com",
    "noJs":   { "status": 200, "title": "…", "wordCount": 412 },
    "render": { "status": 200, "title": "…", "wordCount": 1180 },
    "diff": { "contentOnlyInJs": true },
    "usage": { "cost_usd": 0.0012, "free_usd": 0.0012, "paid_usd": 0 }
  }
}

Guide: SEO Audit API.

Collectors

GET /v1/scraper/collectors catalog, no charge

GET /v1/scraper/collectors/{slug}

POST /v1/scraper/collectors/{slug}/run from $0.0005 / row

GET /v1/scraper/collectors/runs

GET /v1/scraper/collectors/runs/{runId}

Ready-made scrapers on a semantic input (keyword + location, place id, product id, domain) instead of URLs. Each publishes a versioned schema. Billed per delivered row. Long runs return 202 + statusUrl. Export with ?format=csv.

Param	Type	Default	Description
<input fields>	per collector	—	The collector's semantic input — keyword + location, data_id, product_id, domains… — exactly as published in input_schema by GET /v1/scraper/collectors/{slug}.
max_results	integer	per collector	How many rows to deliver at most (clamped to the collector's cap). You pay only for delivered rows.
country / lang	string	—	Proxy exit geo (ISO code) and Google interface language, where the collector supports them.
async	boolean	false	Force background processing: 202 + statusUrl to poll. Long inputs go async automatically.

Request

curl "https://api.quanticdata.io/v1/scraper/collectors" \
  -H "Authorization: Bearer qd_live_YOUR_KEY"

curl -X POST "https://api.quanticdata.io/v1/scraper/collectors/google_maps_places/run" \
  -H "Authorization: Bearer qd_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"query": "pizza restaurants", "location": "Brooklyn, NY", "country": "us", "max_results": 20}'

curl "https://api.quanticdata.io/v1/scraper/collectors/runs/RUN_ID?format=csv" \
  -H "Authorization: Bearer qd_live_YOUR_KEY"

Response

{
  "type": "response",
  "message": "Collector run complete",
  "payload": {
    "run_id": "cmsx…", "slug": "google_maps_places", "status": "done",
    "count": 20, "partial": false,
    "results": [{ "rank": 1, "name": "…", "address": "…", "place_id": "ChIJ…" }],
    "usage": { "cost_usd": 0.02, "free_usd": 0.02, "paid_usd": 0 }
  }
}

Catalog: Collectors.

Proxies

GET /v1/public/proxies

POST /v1/public/proxies/generate

GET /v1/public/proxies/whitelist-ip

POST /v1/public/proxies/whitelist-ip

DELETE /v1/public/proxies/whitelist-ip

Same host and Bearer key. List your plans, then generate ready-to-use endpoint strings (HTTP or SOCKS5). Geo targeting depends on the plan: country / state / city, ISP or ASN. Whitelist is for IP-auth on Residential Basic, Datacenter, ISP, IPv6 and Mobile — Residential Premium and Private use user:pass and skip it. Valid location codes: GET /v1/public/geo/countries?planType=… (also states, cities, asns).

Param	Type	Default	Description
orderId req	string	required	From GET /v1/public/proxies.
protocol	string	http	http \| socks5.
format	string	user:pass@host:port	Also host:port:user:pass, http://user:pass@host:port, socks5://user:pass@host:port.
quantity	integer	10	How many strings to return (max 10,000).
country / state / city	string	—	Geo targeting. state alias: region. Slugs from the geo endpoints.
rotation	string	rotating	rotating \| sticky \| static (IPv6 only).
sessionTime	integer	10	Sticky duration in minutes (1–1440; Basic/Datacenter min 3).
isp / asn	string	—	isp: Premium and Mobile V2. asn: Basic/Datacenter Basic (e.g. AS12345).

Request

curl "https://api.quanticdata.io/v1/public/proxies?active=true" \
  -H "Authorization: Bearer qd_live_YOUR_KEY"

curl -X POST "https://api.quanticdata.io/v1/public/proxies/generate" \
  -H "Authorization: Bearer qd_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"orderId":"YOUR_ORDER_ID","protocol":"http","quantity":10,"country":"us","rotation":"rotating"}'

Errors and rate limits

HTTP status plus {"type":"error","message":"…"}.

Param	Type	Default	Description
400	status	—	Malformed input — the message says which param and why.
401	status	—	Missing, invalid or expired API key.
403	status	—	The endpoint is not enabled for your account, or no proxy plan is available.
429	status	—	Your key's rate limit, or an upstream throttle from the target — back off on Retry-After.
5xx	status	—	Our side (including scraper-service timeouts). Never billed.

PAYG allows 60 requests per minute. Headers: X-RateLimit-Limit, X-RateLimit-Remaining, X-RateLimit-Reset. A 429 also sends Retry-After.

MCP server

npx -y quanticdata-mcp exposes scrape, search, map, crawl, batch, seo_audit, collectors and proxy tools to Claude, Cursor and any MCP client. Tools: MCP server page.

claude mcp add quanticdata \
  -e QUANTICDATA_API_KEY=qd_live_your_key_here \
  -- npx -y quanticdata-mcp

{
  "mcpServers": {
    "quanticdata": {
      "command": "npx",
      "args": ["-y", "quanticdata-mcp"],
      "env": { "QUANTICDATA_API_KEY": "qd_live_your_key_here" }
    }
  }
}

// .cursor/mcp.json
{
  "mcpServers": {
    "quanticdata": {
      "command": "npx",
      "args": ["-y", "quanticdata-mcp"],
      "env": { "QUANTICDATA_API_KEY": "qd_live_your_key_here" }
    }
  }
}

Pricing

List prices on pay-as-you-go. Volume tiers take up to 30% off. You pay the tier that actually ran. Failed calls are never billed. Crawl and batch charge up front and refund the unfetched share.

Call	List price
Scrape (per page)	$0.0002
Scrape, JS render (per page)	$0.001
SERP (per search)	$0.0005
SERP, rendered (per search)	$0.002
Map (per site)	$0.0005
Crawl (per page)	$0.0003
Batch (per URL)	$0.0002
SEO audit (per URL)	$0.0012
Collectors (per delivered result)	from $0.0005

FAQ

How do I authenticate?

Every request carries Authorization: Bearer qd_live_… (or your key prefix). One key covers scrape, SERP, map, crawl, batch, SEO audit, collectors and proxy generation. Get a key at app.quanticdata.io/register — $2 free every month, no card.

What is the base URL?

https://api.quanticdata.io/v1. Short aliases (POST /v1/scrape, /v1/serp, /v1/map, /v1/crawl, /v1/batch, /v1/seo-audit) and the long form (POST /v1/scraper/extract, /v1/scraper/serp, …) hit the same API. Collectors live under /v1/scraper/collectors. Proxy plans live under /v1/public/proxies.

Is there a free web scraping API?

Yes. Every account with a permanent email address includes $2 of free usage each month with no card. Failed calls are never billed. PAYG is 60 requests per minute; subscription tiers raise limits and discount unit prices.

How do I scrape a page in Python?

POST https://api.quanticdata.io/v1/scrape with requests: Authorization Bearer header and a JSON body containing url. Read payload.content for Markdown. The same envelope is used from Node, curl, or any HTTP client.

Do failed requests cost money?

No. Pay per success: a blocked page, a 4xx/5xx from the target, a timeout or a failed job is never charged. Crawl and batch are charged up front on requested volume and the unfetched share is refunded when the job settles. Collectors bill per delivered row — zero rows, zero charge.

How much does a scrape cost?

A successful scrape is $0.0002 per page on the TLS/request tier and $0.001 when the stealth browser renders JavaScript. SERP starts at $0.0005 ($0.002 rendered), map $0.0005 per site, crawl $0.0003 per page, batch $0.0002 per URL, SEO audit $0.0012 per URL, collectors from $0.0005 per delivered result.

What is the difference between scrape, batch and crawl?

Scrape fetches one URL and returns immediately. Batch takes a list of URLs you already know (up to 5,000) and runs them asynchronously. Crawl starts from a seed URL and follows links (up to 500 pages, depth 10). Map lists every URL of a site without fetching page bodies.

How do I use this from Claude or Cursor?

Install the MCP server: npx -y quanticdata-mcp with env QUANTICDATA_API_KEY. Claude Code: claude mcp add quanticdata -e QUANTICDATA_API_KEY=qd_live_… -- npx -y quanticdata-mcp. Cursor reads .cursor/mcp.json. Tools include scrape, search, map, crawl, batch, seo_audit, collectors and proxy tools (list_proxies, generate_proxies).

What does a successful response look like?

Every endpoint returns { "type": "response", "message": "…", "payload": {…} }. On failure, type is "error" and the reason is in message — nothing is billed. Successful calls also carry payload.usage with cost_usd, free_usd and paid_usd.

Are there rate limits?

PAYG keys are limited to 60 requests per minute. Responses include X-RateLimit-Limit, X-RateLimit-Remaining and X-RateLimit-Reset; a 429 also sends Retry-After in seconds. Subscription tiers raise the cap.

Where do I get an API key?

Sign up at https://app.quanticdata.io/register. No card. Keys start with qd_live_. Reveal and rotate them on the API Keys page in the dashboard.

Can I get Markdown instead of HTML?

Yes. format defaults to markdown on scrape, crawl and batch. Pass html or text, or formats as an array to receive several at once. contentMode smart strips nav/footer/cookie chrome; article uses Readability; full keeps the whole body.