Connect Claude to Firecrawl with MCP (Web Scraping Superpower)

Full video transcript

Go to Claude. Go to your settings. Go to connectors. And in connectors, add a custom connector. Name it. Go to docs.firecrawl.dev/mcp-server. Copy the remote hosted URL into remote MCP server URL. And then you copy your API key and then you’re set.

Hey everybody. In today’s video, I’m going to be sharing with you over some coffee in Claude how to connect Claude with Firecrawl to get more relevant data from the companies, the clients, the products, and the services that you’re building — using Firecrawl and their web data API directly with Claude through their MCP. So you can power up your search when you’re using Claude.

If you’re not familiar with Firecrawl, head on over to firecrawl.dev and check out their website. It’s pretty self-explanatory. You can turn websites into LLM-ready data. You can power apps with Firecrawl from any website. It’s open source and it can do a whole lot. And you can connect it directly inside of your Claude account to give entirely new superpowers to Claude’s capabilities of searching and finding actual data that you’re looking for using Firecrawl.

If you’re not familiar with Model Context Protocol (MCP), I’d recommend a quick Google search to catch yourself up on what it means. It’s basically a way to connect apps to communicate with apps, using Claude as the communication factor. So what I’m going to do is go to my settings first and select connectors. You’ll see I have Drive, Gmail, Google, GitHub, Notion, then Zapier. All of these connectors — outside of the Zapier MCP server — are natively integrated inside of Claude right now. But there are thousands of other products you can connect to via MCP inside of Claude that might not have a native connector. You can use Zapier to connect to most of these apps, but then you end up running up your Zapier usage bill, which can get expensive pretty fast. So what I’ve done is set up my Firecrawl MCP server inside of Claude — basically giving Claude a tool to use Firecrawl.

To do that, all you need to do is configure it. When you click configure and you’ve already set it up, it’ll give you access to a number of these different tool permissions: allow, needs approval, blocked. But how do we even get it configured in the first place? If you’re not techy, this can seem confusing. It’s quite simple. The easiest way I found was to literally Google search “how do I find a Firecrawl MCP URL” and then add my API key to it.

What you need to do is put your remote MCP server URL here. There are advanced settings which I haven’t needed to get into. This remote MCP server URL is basically the main URL that Firecrawl will have, and you set it up with your API key at the end of it, then add that to the MCP tool as a connector. So Google search “firecrawl mcp URL” / “firecrawl mcp server.” I’ll link the installation page directly. Literally follow it step for step — it’s very simple. The remote hosting URL is what you want: mcp.firecrawl.dev and then the brackets {firecrawl_api_key}/v2/mcp. At some point they may update this to V3 and you may need to change it, but for now that’s all you need. Copy this URL.

Then log in to your Firecrawl account — you can create one for free — and front and center on your overview is the API key. Copy the key (I’m not going to show mine), paste it in, and click Add. That’s pretty much it. When I tried to add it, it said the server already existed so it wouldn’t duplicate it, which is fine.

It’s as simple as that: go to Claude, go to settings, go to connectors, add a custom connector, name it Firecrawl or whatever you want, go to docs.firecrawl.dev/mcp-server, copy the remote hosted URL into your remote MCP server URL, copy your API key over, and you’re set. Once you do that, you can use it whenever you’re executing skills, recipes, or prompts, and you can track your usage directly inside Firecrawl. You can see I’ve used about 44% of my 500 credits available this month — you can see how credits are pulled and the activity log of how it’s scraping websites.

Overall, I found this to be like giving a superpower to Claude in the form of web research and scraping websites and getting actual relevant data for what I’m looking for. I’ve been using it for building lists and documents and research around events — reaching out to vendors, community leaders, and influencers, doing all sorts of targeted outreach — and it’s using Firecrawl to scrape, find, and source that information. It’s that simple to get it set up and working inside your Claude account. I hope you found this useful. If you did, like, subscribe, and share, and I’ll see you in the next one.