THE WEB SCRAPER API

Get any web page as clean text in 0.4 seconds
Or you don’t pay

Read live, never a stale database No login, no cookies Free to try
Try for free
GET/v1/search?type=webpage&url=https://blog.google/technology/ai/
{
  "text": "AI\n\nThe work we're doing to make AI helpful for everyone.\n\nAI\n\nI/O 2026: Welcome to the agentic Gemini era\n\nThe latest from Google I/O: See how we’re helping you get more done with Gemini.\n\nGemini App\n\nCreate your best tracks yet with Lyria 3.5 in Gemini.\n\nTurn your voice into action with new productivity features in Gemini Live\n\nHere’s how to use intelligent dictation in Gemini for macOS.\n\nResearch\n\n5 amazing visuals show how the male fruit fly’s brain map is advancing neuroscience\n\nAsk a Scientist: How can researchers use AI to predict a flood?\n\nOperation Blue Skies: Reducing aviation climate impact with AI\n\nView more from Google Research\n\nDevelopers\n\nPairing Google Antigravity with Gemini 3.7 Flash solves notable multi-agent math and engineering problems.\n\nGemini Omni 1.1 Flash lets you build with more control\n\nHow developers build AI for good with Gemma 4\n\nView more from Developer tools\n\nAll the Latest",
  "metadata": {
    "title": "Official Google AI news and updates | Google Blog",
    "viewport": "width=device-width, initial-scale=1.0, user-scalable=1.0, minimum-scale=1.0",
    "optimize_experiments": "[]",
    "description": "Explore the cutting-edge work Google is doing in AI and machine learning.",
    "og:type": "website",
    "og:title": "AI",
    "og:description": "Explore the cutting-edge work Google is doing in AI and machine learning.",
    "og:image": "https://blog.google/static/blogv2/images/google-1440x754.jpg?version=pr20260902-1737",
    "og:site_name": "blog.google",
    "og:url": "https://blog.google/innovation-and-ai/technology/ai/",
    "twitter:card": "summary_large_image",
    "twitter:url": "https://blog.google/innovation-and-ai/technology/ai/",
    "twitter:title": "AI",
    "twitter:description": "Explore the cutting-edge work Google is doing in AI and machine learning.",
    "twitter:image:src": "https://blog.google/static/blogv2/images/google-1440x754.jpg?version=pr20260902-1737",
    "twitter:site": "@google",
    "page": "35370",
    "locale": "en-us",
    "published_time": "2018-06-07T18:13:42.706255+00:00",
    "content_type": "blogv2.subcategorypage",
    "tags": "AI",
    "authors": "",
    "gtm-tag": "GTM-TRV24V",
    "canonical": "https://blog.google/innovation-and-ai/technology/ai/"
  },
  "jsonld": null,
  "tookMs": 313
}
· TWO REAL RESPONSES, PULLED LIVE 2026-09-05 ·
The problem

Raw pages waste your model’s tokens

Menus, cookie banners, footers, scripts. Feed a raw page to a model and most of what it reads is filler. Fetch and clean it yourself and you own the parser forever.

1
request per page, text and metadata together
0
parsers you have to keep alive
What that costs youTokens spent on boilerplate. Answers built on the wrong text. A fetch layer you rebuild every quarter.
The fix

A clean read of the page does not

CrustAPI reads the page the moment you ask and hands back the text, the title, description and social tags, and any structured data the page carries, in 0.4 seconds.

The article, not the menus Metadata as fields Structured data included Empty pages are free

You get the page. Cheaper. Fresher. Easier.

Live read : noun

We read the page the moment you ask. Not a copy from last month.

Any page, as text your model can use

0.4 seconds. One bill. Nothing to babysit.

Try for free
The process

Here’s how you get the page you want

STEP 1
1

Your page in one line. Or no code at all

Paste the page address, in the playground or in one line of code.

Zero setup. Zero babysitting.
· 0.4 SECONDS LATER ·
STEP 2
2

The page, cleaned. Right now

The readable text, the title and description, the social tags and any structured data. Ready for your model, your index or your spreadsheet.

FieldWhat comes back
metadata.titleOfficial Google AI news and updates | Google Blog
metadata.descriptionExplore the cutting-edge work Google is doing in AI and machine learning.
metadata.og:imagehttps://blog.google/static/blogv2/images/google-1440x754.jpg?version=pr20260902-1737
textAI The work we're doing to make AI helpful for everyone. AI I/O 2026: Welcome to the age…
+ every other meta and social tag, and any structured data the page carries
The page as it is today.
Text without menus, ads or footersEvery meta tag as a fieldOne price per page, every field included
· PAY ONLY FOR RESULTS · EMPTY PAGES ARE FREE ·
STEP 3
3

Ask for Markdown too. Same line, one more word

Add includeMarkdown and the page comes back with headings and links kept, the shape retrieval systems like best.

Ask forWhat comes back
includeMarkdown=truethe page as Markdown as well
the same pagetoday’s version
webthe results page for any query, to find pages worth reading
+ news, Scholar, patents and 9 more
Text your model can trust.
Try for free
· FREE TO TRY · YOUR BALANCE NEVER EXPIRES ·
Works with your stack
What we do so you don’t

We read the web for you. Live

You get the rows you want. Right today. For less.

PAY ONLY FOR RESULTS. WRONG RESULT, MONEY BACK.
1

We read it live

The moment you ask. Any public page, cleaned to its text, with the metadata and structured data it carries.

2

We keep it working

the web changes its pages all the time. We keep the read working through every change, so you never run, patch or babysit a scraper.

3

We clean it

Menus, banners and footers stripped. Meta and social tags as fields. Structured data parsed, not left as a blob.

4

Guaranteed right

You pay only when data comes back. Empty results cost nothing. Report a wrong result and the money comes back.

BONUS: A data expert on email

Stuck on a field, a filter or a weird result? Email support@crustapi.com and a data expert answers within a day. Not a bot.

Try for free

All without

Proxies or browsersnothing to run, patch, or babysit
A vendor per sourceone account, one bill
Paying for failuresempty results come back free
Bonus

Or let your AI do it

Plug in our MCP server. Your agent does the rest, paying per result.

Claude logo ChatGPT logo MCP
Pricing

How much does it cost?

Pay as you go. One price per page, every field included, and you pay only when the page comes back. Larger deposits get lower rates for those funds until they are spent.

PagesCostRate
1,000$1.00$1.00 per 1,000, pay as you go
49K$49$1.00 per 1,000 with a $49 deposit
200K$149$0.76 per 1,000 with a $149 deposit
980K$549$0.56 per 1,000 with a $549 deposit
4.9M$1,999$0.41 per 1,000 with a $1,999 deposit
19.7M$6,500$0.33 per 1,000 with a $6,500 deposit
98.2M$27,500$0.28 per 1,000 with a $27,500 deposit
192.3M$50,000$0.26 per 1,000 with a $50,000 deposit
500M$100,000$0.20 per 1,000 with a $100,000 deposit
MoreCustomOngoing feeds, managed delivery, or volume below the ladder. Tell us what you need.

$6 free every month, no card. Empty results cost nothing. Your paid balance never expires.

Reference

What you get per page

FieldWhat it holds
textthe readable text of the page
metadatatitle, description, viewport, and every og: and twitter: tag
jsonldthe structured data the page declares
markdownthe page as Markdown, when you ask for it
How to use

How do I use the Web Scraper API?

  1. 1

    Create a free account. No credit card required.

  2. 2

    Paste the page address.

  3. 3

    Add includeMarkdown if you want headings and links kept.

  4. 4

    Download the list as a spreadsheet, use your API key, or tell your AI tool to pull it.

In code it is one line:

curl "https://crustapi.com/v1/search?type=webpage&url=https://blog.google/technology/ai/" \
  -H "x-api-key: YOUR_KEY"
Input

What can I set?

ParameterWhat it doesExample
typewebpage for one page. web for a results page, to find pages worth reading.webpage
urlThe page address.https://blog.google/technology/ai/
includeMarkdownAlso return the page as Markdown.true
FAQ

Common questions

One request that returns any public page as data: the readable text, the title, description and social tags, and any structured data the page declares. You send the address. The page comes back in 0.4 seconds.

That is what it is for. The text comes without menus, ads and footers, and Markdown keeps headings and links, so retrieval and answers run on the article, not the menus.

We read public pages only. Anything behind a login comes back empty, and empty is free.

0.4 seconds. That is the median on customer calls over the two weeks before this page was built.

$1.00 for every 1,000 pages, with lower rates for larger individual deposits. A page that returns nothing costs nothing. $6 free every month, no card.

We read public pages, logged out, and never use accounts or captcha tricks. Respect the page owner’s terms for how you use the text. Public, logged-out collection has repeatedly defeated computer-fraud and contract claims. We are not lawyers.

The CrustAPI guarantee

Get any web page as clean text in 0.4 seconds
Or you don’t pay

Pay only for results Empty results are free Your balance never expires Deposit more, pay less
Try for free
No credit card required.

Every Google endpoint

Search, Maps, reviews and 9 more. One account, one bill.

Public pages only. No logins. No fake accounts.
GDPR COMPLIANTCCPA COMPLIANTSOC 2 IN PROGRESS