/scrape200
A page to Markdown, with real browsers for hard sites.
ADcookies? accept
# news.ycombinator.com1. Baidu released ERNIE 5.5 ...2. Lemmings CSS Animation3. A thousand fewer paths ...
✓ 200 · 4.9 KB Markdown · 2.25 s
The official hosted API from the team behind the open-source crawlerFrom the team behind the open-source crawler
The crawler you already use. Hosted, with one API call.
Turn any URL into clean, LLM-ready Markdown or structured JSON. Scrape, extract, and search with one key. We run the browsers and proxies.Turn any URL into clean, LLM-ready Markdown or JSON. Same engine as the open-source library; we run the browsers and proxies.
US, Singapore & EU regionsGDPR and PDPADPA availableSOC 2 in progress
1,000 free credits every month. No credit card.
Building for a team? Talk to us · Hosted API from the Crawl4AI team1,000 free credits a month. No credit card.
Building for a team? Talk to us · Hosted API from the Crawl4AI team1,000 free credits a month. No credit card.
$ crawl scrape
--instruction
--country
Live · play limits✓ 200 OK
# Type a URL above and hit Run - real output renders here
One key. One bill. Every endpoint.
Scrape, extract, search, watch, whole sites and more, all on the same key. Every call returns its status, its time and its clean result. You pay for what the page needed, and failed calls cost nothing.You pay for what the page needed. Failed calls cost nothing.
/extract200
Structured JSON from an instruction you write in plain words.
pointstitle
172Aluminum foil (2021)
48Kani: A Model Checker for Rust
146AMD Ryzen AI Halo
instruction: titles and points · typed JSON
/recipes
Data that is ready to use.
93 ready-made recipes turn the sites you care about into structured rows: leads, launches, social media, shops and local search.
Copy-paste recipes: docs to Markdown, price watch, lead research, RAG ingest.
LinkedInAmazonGoogle MapsYouTubeRedditShopifyZillowGlassdoorProduct HuntBooking.comLinkedInAmazonGoogle MapsYouTubeRedditShopifyZillowGlassdoorProduct HuntBooking.com
POST /recipes/linkedin-company-peoplePeople at a company on LinkedIn200
| name | headline | location |
|---|---|---|
| A. Rivera | Head of Growth | Austin, US |
| M. Chen | VP Sales | Singapore |
| L. Novak | Data Lead | Berlin, DE |
| S. Okafor | RevOps Manager | London, UK |
example rows · real fields · ready to load into a sheet or a database
POST /recipes/producthunt-todayProduct Hunt: today's launches200
| name | tagline | votes |
|---|---|---|
| Notewise | AI notes for every meeting | 612 |
| Shipkit | Deploy previews in one click | 488 |
| Tallyho | Count anything with your camera | 301 |
| Pagewell | Docs your team will read | 256 |
example rows · real fields · ready to load into a sheet or a database
POST /recipes/instagram-profileProfile of an Instagram account200
| username | followers | posts |
|---|---|---|
| @studio.example | 48,210 | 1,204 |
| @coffee.example | 12,880 | 640 |
| @travel.example | 231,004 | 3,912 |
| @shop.example | 7,415 | 288 |
example rows · real fields · ready to load into a sheet or a database
POST /recipes/google-maps-placesGoogle Maps business search200
| # | name | category |
|---|---|---|
| 1 | Corner Café | Coffee shop |
| 2 | Daily Grind | Coffee shop |
| 3 | Bean There | Café |
| 4 | Roast Room | Coffee roaster |
example rows · real fields · ready to load into a sheet or a database
POST /recipes/amazon-searchAmazon search results200
| title | price | rating |
|---|---|---|
| Wireless earbuds, 30 h battery | $39.99 | 4.5 |
| Noise-cancelling headphones | $89.00 | 4.6 |
| USB-C fast charger, 65 W | $24.99 | 4.7 |
| Laptop stand, aluminium | $29.50 | 4.4 |
example rows · real fields · ready to load into a sheet or a database
/serp200
Search result pages as data.
- The Best CRM Software You Should Considerneilpatel.com
- What is the best CRM software on the market?g2.com
- Top 10 Best CRM Software & Systemsem360tech.com
- The Best CRM Software 2024theaccessgroup.com
✓ 10 results · 5 blocks · 1.18 s
/price
Plan prices from pricing pages, in dollars.
HubspotStarter£7 /mo$9.43
Professional£81 /mo$109.12
CanvaPro£100 /yr$134.71
✓ 3 sites · 3 with a price · 24.8 s
/watch
Watch a page is our change monitor. Say what to look for, and we alert you when it changes.
example.com/pricingPro $49$59
Changed Pro plan: $49 → $59 example
$1.00 per check · $3.00 per change
/scan · /sitewhole sites
Go deep into a whole site.
/scan lists every URL of a site from its sitemaps, its home page and public crawl indexes, in one call, in 1 to 3 seconds for most sites.
/site crawls a site level by level from a start page, only inside the site, and returns every page as Markdown, as a job.
curl https://api.crawl4ai.com/site \
-d '{"url":"https://docs.crawl4ai.com/","depth":2}'depth 2 · up to 100 pages by default · example
Results captured from the live API, 5 Nov 2026.
Credits per call: plain page 0.1, real browser 0.5, hard site 2, search 1.
Batch up to 50 URLs, jobs up to 10,000 URLs, and page-change monitors.
Batch up to 50 URLs, jobs up to 10,000 URLs, and page-change monitors.
Get a key, then connect any of these.1,000 free credits every month. No credit card.
Get a free API key
/recipes
Browse 93 recipes
Data that is ready to use.
93 recipes turn the sites you care about into rows: leads, launches, social, shops, local search.
LeadsLaunchesSocialLocal + SEOShops
linkedin-company-people200
A. RiveraHead of Growth
Austin, USM. ChenVP Sales
SingaporeL. NovakData Lead
Berlin, DEYour agent can set this up by itself.
Point any coding agent at our llms.txt, or add the hosted MCP server. The agent reads the setup, gets a key approved, and starts calling tools. No SDK to install.Paste one line into Claude Code, Cursor, Codex or any agent that reads a URL.
- Claude Code
- Cursor
- Codex
- Windsurf
- OpenCode
- Gemini CLI
- ChatGPT
- Claude Desktop
- OpenClaw
MCP tools: scrape, search, serp, answer, extract, batch, jobs, balance, estimate, top-up and page watchers.
Paste this into your agentPromptprompt
set up crawl4ai — https://crawl4ai.com/llms.txt
What the agent readsRead llms.txt
Crawl4AI Cloud is the official hosted API of the open-source Crawl4AI project. API base: https://api.crawl4ai.com Auth: Authorization: Bearer sk_live_… (or x-api-key) MCP (streamable HTTP): https://api.crawl4ai.com/mcp MCP auth: sign in and approve access, or add a key. Not affiliated: crawl4ai.org
One engine, run for you.
Crawl4AI is the open-source crawler we wrote. Cloud runs it for you, with everything a production crawler needs around it.
Cloud
export CRAWL4AI_KEY=
curl api.crawl4ai.com/scrape -d '{"url":"…"}'
- Scrape, search, SERP, extract, recipes, price and page watchers on one key.
- Hosted MCP server for agents. Batch and job endpoints for scale.
Browsers and proxies, run for you
Real browsers, residential and ISP proxies, scaled by us.
Hard sites handled
We keep up with anti-bot changes. A failed call costs nothing.
Three regions, every call traced
US, EU and Singapore. Each answer says what it took.
Ready for your security review
Roles and an audit log from Builder; SSO and an SLA on Enterprise.
Have a team that runs browsers, proxies and servers? The open-source library is on GitHub, and you can host it yourself. Self-hosting guide →
What does running it yourself really cost? See build vs buy →
Start free. Pay as you grow.
Credits are shared across every endpoint. A plain page is 0.1 credit, a real browser page 0.5, a hard site 2. Failed calls are free.
Free
$0freeno card needed
1,000credits a monthup to 10,000 pages
- Rate
- 10 / min
- At once
- 2
- Best for
- prototypes
- Support
- Discord community
Starter
$10 a month
10,000credits a monthup to 100,000 pages
- Rate
- 120 / min
- At once
- 10
- Extra 1k credits
- $1
- Support
- email, 2 business days
BuilderMost popular
$99 a month
125,000credits a monthup to 1,250,000 pages
- Rate
- 600 / min
- At once
- 20
- Extra 1k credits
- $0.79
- Support
- email, 1 business day
Scale
$499 a month
750,000credits a monthup to 7,500,000 pages
- Rate
- 1,500 / min
- At once
- 50
- Extra 1k credits
- $0.67
- Support
- priority email, 4 business hours
Enterprise
Customlet’s talkbuilt around your volume
Customcredits a monthpages to match your volume
- Rate
- custom
- Fetching
- US, EU or Singapore
- Terms
- SLA, DPA
- Sign-in
- SSO (SAML, OIDC)
- Support
- named engineer, SLA by severity
Page counts assume plain pages. A page that needs a real browser is 0.5 credit, so a plan gives five times fewer of those. Yearly prices are shown per month; Starter bills $100 a year, Builder $990, Scale $4,990. Failed calls cost nothing.
What your plan buys, here and elsewhere.
Each number is how many you get for that plan’s monthly price, against each seller’s plan at a similar price. Our credits are shared, so spend them on any row in any mix.
| Crawl4AIStarter · $10 a month | FirecrawlHobby · $19 | ScrapingBeeHobby · $19 | SerpApiStarter · $25 | TavilyProject · $30 | ScraperAPIHobby · $49 | |
|---|---|---|---|---|---|---|
| Plain page | 100,000 | – | 75,000 | – | 20,000 | 100,000 |
| Page in a real browser | 20,000 | 5,000 | 15,000 | – | 10,000 | 10,000 |
| Hard site (anti-bot) | 5,000 | 5,000 | 3,000 | – | not stated | 4,000 |
| Web search | 10,000 | 2,500 | 7,500 | – | 4,000 | – |
| Google results | 10,000 | – | 7,500 | 1,000 | – | 4,000 |
| Structured JSON (extract) | ~6,600 | 1,000 | 7,500 | – | research only | 100,000 |
| Ready data recipes | 93 sites | – | – | – | – | a few sites |
| Plan prices from pricing pages | 2,000 sites | – | – | – | – | – |
| Watch a page for changes | 10,000 checks | 5,000 checks | – | – | – | – |
| A whole site, every page | 100,000 pages | yes | – | – | yes | yes |
| Failed calls | free | partly charged | partly charged | free | free | partly charged |
– = not offered. From each seller’s own pricing and docs pages, checked 6 November 2026. Other sellers shown at their cheapest paid plan.
| Crawl4AIBuilder · $99 a month | ScrapingBeeStartup · $99 | FirecrawlStandard · $99 | TavilyBootstrap · $100 | SerpApiDeveloper · $75 | ScraperAPIStartup · $149 | |
|---|---|---|---|---|---|---|
| Plain page | 1,250,000 | 1,000,000 | – | 75,000 | – | 1,000,000 |
| Page in a real browser | 250,000 | 200,000 | 100,000 | 37,500 | – | 100,000 |
| Hard site (anti-bot) | 62,500 | 40,000 | 100,000 | not stated | – | 40,000 |
| Web search | 125,000 | 100,000 | 50,000 | 15,000 | – | – |
| Google results | 125,000 | 100,000 | – | – | 5,000 | 40,000 |
| Structured JSON (extract) | ~83,000 | 100,000 | 20,000 | research only | – | 1,000,000 |
| Ready data recipes | 93 sites | – | – | – | – | a few sites |
| Plan prices from pricing pages | 25,000 sites | – | – | – | – | – |
| Watch a page for changes | 125,000 checks | – | 100,000 checks | – | – | – |
| A whole site, every page | 1,250,000 pages | – | yes | yes | – | yes |
| Failed calls | free | partly charged | partly charged | free | free | partly charged |
– = not offered. From each seller’s own pricing and docs pages, checked 6 November 2026. Other sellers shown at their plan closest to $100 a month.
| Crawl4AIScale · $499 a month | ScrapingBeeBusiness+ · $599 | FirecrawlGrowth · $399 | TavilyGrowth · $500 | SerpApiBig Data · $275 | ScraperAPIScaling · $475 | |
|---|---|---|---|---|---|---|
| Plain page | 7,500,000 | 8,000,000 | – | 500,000 | – | 5,000,000 |
| Page in a real browser | 1,500,000 | 1,600,000 | 500,000 | 250,000 | – | 500,000 |
| Hard site (anti-bot) | 375,000 | 320,000 | 500,000 | not stated | – | 200,000 |
| Web search | 750,000 | 800,000 | 250,000 | 100,000 | – | – |
| Google results | 750,000 | 800,000 | – | – | 30,000 | 200,000 |
| Structured JSON (extract) | ~500,000 | 800,000 | 100,000 | research only | – | 5,000,000 |
| Ready data recipes | 93 sites | – | – | – | – | a few sites |
| Plan prices from pricing pages | 150,000 sites | – | – | – | – | – |
| Watch a page for changes | 750,000 checks | – | 500,000 checks | – | – | – |
| A whole site, every page | 7,500,000 pages | – | yes | yes | – | yes |
| Failed calls | free | partly charged | partly charged | free | free | partly charged |
– = not offered. From each seller’s own pricing and docs pages, checked 6 November 2026. Other sellers shown at their plan closest to $500 a month.
Built to pass your security review.
Large teams run Crawl4AI Cloud on custom plans. Everything below is available on Enterprise; the rest is a 30-minute call.Everything below is available on Enterprise; the rest is a 30-minute call.
- SSO, roles, audit log
- US, EU or Singapore
- DPA, invoice billing, SLA
- Support from the team that writes the crawler
- Custom volume and rate limits above Scale
- Single sign-on (SAML or OIDC) with Okta, Microsoft Entra ID or Google Workspace, roles for every member, and an audit log of every change
- Fetching in the US, EU or Singapore; account data in the EU
- Security review, subprocessor list and DPA on request
- Invoice billing and SLA terms
- Direct support from the team that writes the crawler
Encrypted
TLS 1.2 and 1.3 in transit. Backups use AES-256.
Keys protected
Gates check API keys by their SHA-256 hash. Card numbers stay with Stripe.
Where data lives
Account, usage and billing records in the EU. Browsers run in Germany, Singapore and the US.
Your content
We do not sell data and do not train models on request content.
Retention
Service logs 90 days. Usage and support records 13 months.
Recovery
Daily and weekly backups, restores tested every 6 months. RTO 4 hours, RPO 24 hours.
Access control
MFA for staff and quarterly access reviews.
Vulnerability reports
Reply within 3 business days. Fix targets: critical 7 days, high 30, medium 90.
Fair use
The acceptable use policy asks users to respect robots directives.
GDPRSingapore PDPADPA on requestSOC 2 with an independent auditor: in progress
Subprocessors: Hetzner, AWS, Stripe, Google Workspace, GitHub, OpenRouter, Google Gemini API, Serper, Bright Data, Massive/IPRoyal. Full list and report address on the Security page.
Docs and resources
Reference for every endpoint, plus the places we publish our own numbers.
API docs — every endpoint, auth, limits and errors. Also as llms.txt for agents.DocsChangelog — Cloud and library release notes. Library 0.9.5 shipped 14 Oct 2026.ReleasesStatus — uptime and incidents for the Cloud API.Status pageCrawl4AI vs alternatives — comparison with Firecrawl and others, with sources and a date.CompareRecipes — docs to Markdown, price watch, lead research, RAG ingest.RecipesWeb scraping for RAG — a site to Markdown chunks, and only the changed pages on the next run.Use caseCompetitor price monitoring — pricing pages as rows, change alerts, Google ranks.Use caseBuild vs buy — what a self-hosted crawler costs a month, with sourced numbers and a calculator.Cost
Questions
The ones we get most. Anything else: hello@crawl4ai.com.
Yes. Crawl4AI Cloud is run by the team that writes the open-source Crawl4AI crawler. It runs the same engine, hosted and operated by us.
No. The site crawl4ai.org is not affiliated with Crawl4AI. Use crawl4ai.com, docs.crawl4ai.com, and the GitHub repository.
Yes. The library stays open source and free. When you self-host, you run the browsers, buy the proxies and keep up with anti-bot changes yourself. Cloud does that work, and a failed call costs nothing.
Through the hosted MCP server or the REST API. Setup is in llms.txt. Keep the key in an environment variable.
Calls return a 402 error when credits run out or a spend cap is reached. Buy extra credits at any time. Extra credits never expire. Monthly plan credits do not carry over. Failed requests are not charged.
No. We do not use request content to train models, and we do not sell data.
Within 14 days of your first plan payment or an extra-credit purchase, if you used none of the credits.
For developers
Get a free API key and run your first crawl.
1,000 credits every month. No credit card. Your self-hosted setup keeps working.
For teams
Talk to us about volume, regions and an SLA.
A 30-minute call with the team that writes the crawler. DPA on request, invoice billing, your choice of US, EU or Singapore.