FREE WEB SCRAPING TOOL · NO SIGNUP TO TRY

You shouldn't have to run Chrome to read a webpage

Fetch the raw HTML of any public page — including the JavaScript-rendered ones your own fetch can't reach. Then take the same call to production with one key.

100 credits/mo with a free key1B+ pages fetched

THE TOOL IS THE API

Whatever you type above, you can run in your terminal

The box above is the same fetch — same renderer, same proxies, same cache. With a key you get the v3.0 defaults on top.

Get a free API key Read the docs

Automate this with the API

curl "https://opengraph.io/api/3.0/scrape/https%3A%2F%2Fexample.com?app_id=YOUR_API_KEY"

v3.0 turns on auto_render, auto_proxy and retry by default. Get a free API key →

The HTML your fetch() didn't get

Plenty of the web renders client-side, and plenty of it blocks datacenter IPs. One GET request handles both, and we maintain the browsers and proxies instead of you.

auto_render

No headless browser to babysit

We detect when a page needs JS and render it. Nothing to deploy, no Chromium in your Lambda, no memory ceiling.

auto_proxy

403s become 200s

Millions of IPs across dozens of ISPs, routed per host. Never buy a proxy plan or debug a ban wave again.

retry: true

Failures are on our tab

Retries are built in, and we try not to charge for requests that fail outright — 500s, DNS timeouts, dead hosts.

cache: 24h

Re-reads cost 1 credit

Monitoring the same 500 URLs daily doesn't cost 500 renders. Cache hits are always a single credit.

/markdown

Clean Markdown for RAG

Skip the HTML entirely when you're feeding an LLM. Same key, one endpoint over.

concurrency

Built for the batch job

Fan out thousands of URLs on your own schedule. Concurrency scales with your plan — 1 on the free key, up to 25 on Production — with no rate-limit choreography on your side.

MCP SERVER

Or let your agent scrape it

Point Claude, Cursor or Windsurf at one URL and your agent gets the same scrape and screenshot endpoints this page uses, plus Open Graph data and freeform site queries. Same account, same credits.

{ "mcpServers": { "opengraph": { "url": "https://mcp.opengraph.io/mcp" } } }

MCP docs

What people build with it

Four jobs that start as a one-off scrape and end as a scheduled pipeline.

Competitor price watch

Storefronts render prices in JS and block datacenter IPs. Both handled in one call.

LLM & RAG ingestion

Feed clean Markdown to your index instead of writing another parser per site.

SEO & meta audits

Pull tags, canonicals and OG data across thousands of URLs on a schedule.

Link previews & unfurls

The original job: one endpoint that returns a usable card for any pasted URL.

Questions devs actually ask

Still stuck? The docs have runnable examples for every parameter.

Is the free tool really free, with no account?

Yes — one free scrape, no signup. After that, a free API key gives you 100 credits a month and the full HTML through the same endpoint.

What is a credit, exactly?

One plain fetch is 1. A page that needs browser rendering is 11. Proxy tiers add 10 to 30 depending on the host. Cache hits are always 1.

Why did this site return a 403?

The free tool doesn't use proxies. With a key, auto_proxy routes the request through the cheapest tier that works for that host.

Does it run JavaScript?

When the page needs it. We detect client-side rendering and run a real browser, so SPAs come back fully populated.

Is scraping legal?

Public pages are generally fine to fetch; what you do with the data is on you. Respect robots.txt, terms of service and personal data law.

Scrape one page above. Then scrape a hundred thousand.

100 credits a month, free, no card. Paid plans start at $25 when you outgrow it.

Get my free API key See pricing

Free Tools

More free tools

Everything you need to inspect, scrape, and convert the web — no signup required.