Firecrawl
Run a scheduled AI agent on your Firecrawl — unattended.
Give your Clourou routines Firecrawl and the open web becomes something they can actually read at scale. Firecrawl turns pages — including JavaScript-rendered ones — into clean markdown built for LLMs, maps out a site's URLs, crawls whole sections, runs web searches, and extracts structured data. Where a single-page fetch ends, Firecrawl's crawler begins.
What you can automate
- Docs & changelog watch: Crawl a competitor's docs or changelog section every week and get a summary of what they shipped.
- Site-to-markdown snapshots: Keep a clean markdown copy of chosen site sections in My Files, refreshed on a schedule, for your other routines to read.
- Search-powered briefs: Run a web search and pull the top results as readable markdown for a morning digest.
- Structured extraction: Pull specific fields off pages into tables — prices, specs, dates — without writing a selector.
- URL map & coverage checks: Map a site's URLs on a schedule and flag pages that appeared or disappeared since the last run.
Connecting
On the Connections page, add Firecrawl and paste your API key (firecrawl.dev → API Keys). Scrapes, crawls, and searches spend credits from your own Firecrawl plan — crawls especially, since every page in a crawl counts — so the templates below cap pages per run. One thing to know: the key is checked when tools run, not when you connect, so a mistyped key still "connects" fine and only fails on the first scrape.
Routine templates using Firecrawl
Every Friday, Firecrawl crawls your competitors' docs and changelog sections and you get a plain-English summary of what they shipped this week.
Every Sunday night, Firecrawl crawls the site sections you choose and saves each page as clean markdown in My Files — a fresh, agent-readable knowledge base by Monday morning.
Frequently asked questions
Yes. Attach your Firecrawl connection to a routine and the agent scrapes, maps, or crawls the URLs you name on whatever cadence you set, turning pages into clean markdown each run. Crawls count every page against your Firecrawl credits, so it's worth capping pages per run.
Yes. The agent can crawl the pages, read the markdown, and write up what changed or what's new, then deliver that in-app and by email if you opt the routine into delivery.
A Firecrawl API key, pasted on the Connections page. The key is stored encrypted and used only when your routines run; scrapes and crawls spend credits from your own Firecrawl plan.
Yes. Once scheduled, the routine runs on its cadence with no action from you, and results land in-app and by email. Runs are pay-as-you-go with no subscription.
Yes, within what Firecrawl actually supports — it can pull fields like prices, specs, or dates into structured output without hand-writing selectors. Sites that block bots, gate content behind logins, or change layout often can still cause misses, so treat extracted data as a draft to review.
Automate Firecrawl on a schedule
Connect it to Clourou and attach it to a routine in minutes.
Connect Firecrawl