EU-native web scrapingfor AI agents & developers
turn any webpage into clean markdown, screenshots, metadata, links and html. headless Chrome, rotating proxies and retries built in. processed fully within EU infrastructure.
no credit card · 750 free credits a month · cancel anytime
try it
€0 for a failed scrape€0 for a fully cached repeat1/3/5 credits for http/browser/screenshot
five outputs from one call
markdown, screenshots, metadata, links and html from a single request. no per-site parsers, one response shape for sync and async.
markdown
llm-readyclean content with ads and cookie banners removed. drops straight into a rag pipeline without a parser.
# pricing — acme.dev
acme is a billing platform for usage-based saas.
- starter: €0 /mo
- scale: €49 /mo
[read the docs](https://acme.dev/docs)
screenshots
viewport or full page, on desktop or mobile, as a hosted image url.
metadata
title, description, OpenGraph, canonical, and language, normalized into one block.
{
"title": "acme — billing",
"og:image": "…/og.png",
"lang": "en"
}
links + sitemaps
every discovered link per scrape, plus a /map endpoint for whole-site discovery.
acme.dev
├── /pricing
├── /docs
└── /blog + 214 more
html
cleaned html with scripts, ads and cookie banners removed, or the raw html as the site served it, when you want to parse it yourself.
<article>
<h1>pricing</h1>
<p>acme is a billing platform…</p>
</article>
// cleaned, or raw as served
run everything in the EU
the fetch, the render, the cache, and your stored result never leave EU servers. the proxy exit is the one hop you choose. pick an EU exit and an EU site, and nothing leaves at all.*
everything inside the EUonly the proxy hop or the target site is outside
* processing and long-term storage run on EU servers. screenshots, if you ask for them, are delivered through a content delivery network that can serve them from outside the EU.
no transfer review to open for the scrape path. crawlbrulee is an EU-registered company and your processor, and our data processing agreement is public.
data residency
- company
- EU
- hosting
- EU end-to-end
- processing
- EU end-to-end
- proxy exit
- EU default · worldwide available
- retention
- results deleted within 90 days
- zero data retention
- enterprise · on request
- dpa
- public
- sub-processors
- public
- personal data
- how we handle it
iso 27001 and soc 2 are on the roadmap. ask us if a specific control matters for your review.
built for agents
the interfaces and outputs a coding agent needs, without a wrapper around the api.
plug it into your agent.
native scrape and map tools for Claude Code, Cursor, and Codex, and drop-in skills that teach any coding agent when and how to call crawlbrulee.
mcp server
$ claude mcp add crawlbrulee -- npx -y @crawlbrulee/mcpagent skills
$ npx skills add crawlbrulee/crawlbrulee-skillsscreenshots your model can read.
ask for slices and a full-page screenshot comes back as a stack of tiles sized for an image model. no oversized images to crop yourself.
are you an ai agent?
one markdown file covers what crawlbrulee does, which interface to pick, and how to authenticate.
$ curl https://crawlbrulee.com/.well-known/agent-onboarding.mdthe hard parts, handled
you send a url. the pipeline tries the fastest, most economical option first and steps up to slower, more reliable ones only when a page needs it.
fast on easy pages. stubborn on hard ones. first success wins.
- headless Chrome
- a real browser for javascript pages and screenshots, used only when a page needs it.
- proxies
- basic and advanced tiers with country selection. pick an EU exit with one field.
- auto retries
- on auto we try the light, low-cost proxy first. if a site blocks it, we retry through the stronger proxy tier by ourselves, and you pay the rate of the one that got through.
- clean pages
- ads, trackers, and cookie banners removed before extraction.
- cache
- repeat calls for the same url come back from cache, faster and for 0 credits, inside the freshness window you set.
- async + webhooks
- queue a job, poll it, or get a signed webhook when it is done.
- EU hosting
- EU company, EU servers, public data processing agreement and sub-processors list.
- retention
- results deleted within 90 days, every copy included. with zero data retention, a result stays out of the shared cache and anything we must store is kept for 24 hours, then deleted.
- organizations & roles
- organization management with team, invites, roles, billing, and api keys. managed from the dashboard, on every paid plan. how organizations work.
one api, every interface
rest with the same response shape for sync and async. typed sdks, a cli and an n8n node sit on top of it, or generate a client from the OpenAPI spec.
curl -X POST "https://api.crawlbrulee.com/api/scrape" \-H "Authorization: Bearer $CRAWLBRULEE_API_KEY" \-H "Content-Type: application/json" \-d '{"url": "https://example.com","extract": { "markdown": true, "links": true }}'
// response, same shape in sync and async
{
"markdown": "# example domain\n…",
"links": ["https://www.iana.org/domains/example"],
"metadata": { "title": "Example Domain", … }
}
sdk
sdk docstyped js and Python clients. auth, retries, and error mapping built in.
$ npm i @crawlbrulee/sdkcli
cli docsscrape and map from your terminal or a cron job. clean json when piped.
$ npx crawlbrulee scrape https://example.comopenapi
api referencea public spec for the whole api. generate a typed client in any language.
$ npx openapi-typescript https://crawlbrulee.com/docs/openapi.jsonno-code
crawlbrulee on n8nthe verified n8n node: scrape, map, queued jobs and the scrape-complete trigger, from the node picker.
start free. scale later
pay for outcomes, not attempts. 750 free credits a month, no card.
free
for first experiments and prototypes.
€0/month
no card required
- 750
- credits / month
- 2
- concurrent requests
starter
for side projects and first integrations.
€29/month
or €24 /month on the yearly plan (€290 /year)
- 20,000
- credits / month
- 5
- concurrent requests
growth
for products with steady traffic.
€79/month
or €65 /month on the yearly plan (€790 /year)
- 65,000
- credits / month
- 10
- concurrent requests
pro
most popularfor shipping production workloads.
€149/month
or €124 /month on the yearly plan (€1,490 /year)
- 150,000
- credits / month
- 25
- concurrent requests
ultra
for heavy extraction pipelines.
€399/month
or €332 /month on the yearly plan (€3,990 /year)
- 500,000
- credits / month
- 50
- concurrent requests
€0 for a failed scrape · €0 for a fully cached repeat · 1/3/5 credits for http/browser/screenshot · all prices exclude VAT
enterprise
for large pipelines and regulated industries.
- custom credit quotas
- higher concurrency caps
- zero data retention
- signed DPA
- shared Slack channel
- invoice billing
common questions
short answers here. the long ones are in the docs.
power your next idea
six things you can build with the five outputs.
- rag ingestion
- fresh pages as clean markdown, straight into your index. no parser per site.
- change monitoring
- scrape on a schedule and diff the result. a page that answers 404 still comes back, with its status.
- lead and catalog enrichment
- metadata, links and a screenshot per url, in one call or as queued jobs with a webhook.
- price and stock tracking
- a real browser for javascript shops, a local proxy exit with one field, and repeats inside your freshness window for 0 credits.
- research agents
- scrape and map as native tools over mcp. your agent reads pages as markdown and tall pages as screenshot slices.
- seo and content audits
- title, description, canonical, language and every link per page, with /map to list the whole site first.
ready when you are
start on the free plan: 750 credits every month, no card. scale only when your usage proves the value.
no credit card required · hosted in the EU · cancel anytime