How to Set Up and Use Firecrawl MCP in Cursor

Ninad PathakNinad Pathak
Nov 12, 2026 (updated)

Firecrawl MCP gives your Cursor AI agent web search, scraping, crawling, and structured extraction capabilities without leaving the editor.

When Andrej Karpathy coined "vibe coding" in February 2025, he described building apps in Cursor by just talking, accepting all changes, and never reading diffs.

The tweet got 6.7 million views but at the time, Cursor's AI agent could only work with what was already in your project. It couldn't look up documentation, check a live API, or pull data from the web.

But with the introduction of the Model Context Protocol (MCP), AI coding agents could use external tools through a standardized interface. Firecrawl MCP is one of the most popular MCP servers for Cursor, with 6,500+ stars on GitHub.

I'll walk you through the process to set up Firecrawl in Cursor and use it in a real project.

P.S: Firecrawl is now an official Cursor plugin. Let Cursor agents tap into the live web for real-time context, whenever they need it. Full CLI capabilities (search, scrape, crawl, and browse), industry-leading >80% web coverage, one-click install — get it from the Cursor Marketplace.

Firecrawl Cursor plugin demo

TL;DR

  • The Firecrawl MCP server gives your Cursor AI agent real-time access to the web, allowing it to scrape URLs, search for documentation, and extract structured data while you code.
  • It enables workflows like auto-generating code from live docs, building price trackers, or cloning designs by simply describing what you want in Agent mode (Cmd+I).
  • Setup takes 2 minutes: Add the hosted server URL to Cursor's MCP settings, then sign in with your free Firecrawl account (or use an API key). Nothing to install locally.
  • You don't need to pick tools manually. Just tell the agent "Scrape this URL" or "Search for this error," and it will select the appropriate Firecrawl function.
  • Prefer one-click install? Firecrawl is now an official Cursor plugin — install it directly from the Cursor Marketplace with no JSON config needed.

What is the Firecrawl MCP server?

Firecrawl is the web data API to search, scrape, and interact with the web at scale, turning websites into clean, LLM-ready markdown or structured JSON. It handles JavaScript rendering and dynamic content out of the box. The Firecrawl MCP server wraps this API so any MCP-compatible client (Cursor, Claude Desktop, VS Code, Windsurf) can use it.

The MCP server gives your Cursor AI agent access to 3,000+ tools through Alexandria's data providers, plus core web capabilities:

  • firecrawl_scrape: Scrapes a single URL to clean markdown, HTML, screenshots, or JSON.
  • firecrawl_search: Runs web searches filtered by language, country, and source.
  • firecrawl_developer_search: Queries the Firecrawl Developer Index, a semantic index over 70M+ developer artifacts (READMEs, GitHub issues, pull requests, OpenAPI specs, external docs). The HTTP API also supports filters for types, repos, language, and min_stars. Ideal for finding the exact PR that fixed a bug or the doc page that answers a question from inside Cursor.
  • firecrawl_crawl: Asynchronously follows links across a site respecting depth limits.
  • firecrawl_map: Fast discovery of all URLs on a site without extracting content.
  • firecrawl_extract: Pulls structured data using LLM-powered extraction and a JSON schema.
  • firecrawl_deep_research: Autonomously explores multiple sources for comprehensive analysis.
  • firecrawl_agent: Browses and navigates autonomously to extract data.
  • firecrawl_batch_scrape: Processes multiple URLs in parallel with built-in rate limiting.
  • firecrawl_interact: Scrape a page and immediately take actions in it — click buttons, fill forms, navigate deeper — using natural language prompts or Playwright code. Built on the new /interact endpoint.

Get your free Firecrawl API Key

How to use Firecrawl MCP in Cursor

Once Firecrawl MCP is connected to Cursor, you simply describe what you need in Cursor's Agent mode. The agent automatically picks the right tool. Here are 8 ways you can call the Firecrawl MCP while using Cursor:

1. Reading documentation while building

Suppose you're integrating Stripe and can't remember the webhook verification flow. Instead of switching to a browser, you type:

"Use Firecrawl to scrape the Stripe webhook documentation at stripe.com/docs/webhooks and show me the signature verification steps."

The agent calls firecrawl_scrape, gets clean markdown, and uses it to write your verification code. Because Firecrawl strips navigation, ads, and boilerplate, the agent gets only the relevant content. This means fewer wasted tokens and more accurate code generation.

2. Searching for code patterns

If you need to implement rate limiting in FastAPI but aren't sure about the current best practice, you can say:

"Search for FastAPI rate limiting middleware 2025 and summarize the top approaches."

The agent runs firecrawl_search, reads the results, and gives you a comparison with code snippets. The official docs recommend a two-step workflow here: search first without scrape options (faster), then selectively scrape the most relevant results to get the full content.

3. Extracting structured data for your app

If you want to build a price comparison feature but don't know the exact details, you can let the Cursor agent use Firecrawl to fetch accurate information.

"Extract the product name, price, and availability from these three URLs" (plus a JSON schema).

The agent uses firecrawl_extract and hands you structured JSON you can plug straight into your data model. No regex. No CSS selectors. The LLM interprets the page content against your schema and returns exactly the fields you asked for.

4. Cloning a design

This use case went viral. You can point Firecrawl at any website, scrape it with the branding format (which extracts colors, fonts, typography, and spacing), and then ask Cursor to rebuild the design.

The firecrawl_scrape tool with formats: ["markdown", "branding"] returns both the content and the design tokens. One developer demonstrated cloning an entire website's design in under five minutes using this approach with Cursor Compose.

5. Competitive research during development

"Use deep research to analyze how the top 5 project management tools handle recurring tasks."

The firecrawl_deep_research tool autonomously crawls multiple sites, follows relevant links, and produces a synthesized report. You get a detailed competitive analysis without opening a single browser tab. The findings land directly in your Cursor context where the agent can act on them immediately.

6. Mapping a site before a migration

"Map all URLs on docs.example.com and find pages related to authentication."

The firecrawl_map tool with a search filter returns a URL list you can use to plan a documentation migration or content audit. It's the same first step in a full website migration — mapping every URL before touching the new CMS.

7. Ingesting an entire codebase's docs

Working with an unfamiliar framework?

"Crawl the first 20 pages of docs.astro.build and summarize the routing system."

The agent uses firecrawl_crawl with a low limit to stay within token bounds, then synthesizes what it finds. Keep maxDepth at 2 or 3 to avoid pulling in hundreds of pages. For larger doc sites, use firecrawl_map first to discover URLs, filter to the ones you care about, then firecrawl_batch_scrape only those.

8. Interacting with dynamic pages

Some data is behind a search form, pagination, or login. Instead of writing browser automation scripts, you can use the /interact endpoint to scrape a page and then take actions in it:

"Scrape amazon.com, then search for 'wireless keyboard' and extract the prices of the top 5 results."

The agent scrapes the page to get a scrapeId, then calls firecrawl_interact with natural language prompts to search, click, and extract — all within the same browser session. You can also pass Playwright code for precise control. Sessions support persistent profiles so you can stay logged in across scrapes.

Step-by-step setup guide

Fortunately, it's extremely easy to set up Firecrawl MCP in Cursor. Cursor connects to Firecrawl's hosted MCP server, so there's nothing to install or run on your machine.

Prerequisites

If you'd rather skip the manual config entirely, Firecrawl is available as an official Cursor plugin: one click in the Cursor Marketplace and you're done, no JSON required.

For the MCP server setup, you need two things before starting:

  1. A Firecrawl account. Go to firecrawl.dev and create a free account (1,000 credits per month, no card required). You'll sign in with it from Cursor, or create an API key if you'd rather paste a key into your config.
  2. Cursor latest version. Ensure you are on Cursor 1.0+ (late 2025 onward).

Step 1: Configure the Firecrawl MCP in Cursor

There are three ways to configure Firecrawl MCP in Cursor. Pick whichever matches your version.

Method 1: Edit mcp.json directly

Cursor stores MCP configuration in a JSON file. You can set it globally (~/.cursor/mcp.json) or per-project (.cursor/mcp.json).

Paste this configuration:

{
  "mcpServers": {
    "firecrawl": {
      "url": "https://mcp.firecrawl.dev/v2/mcp-oauth"
    }
  }
}

Save the file, then complete the browser sign-in from Cursor. Cursor opens a window where you sign in to Firecrawl and approve a team. No API key or headers are needed, and you can review and revoke connections from MCP settings.

https://mcp.firecrawl.dev/v2/mcp-oauth is a server URL for Cursor, not a page to open in your browser.

Prefer an API key over browser sign-in? This is the better fit for unattended runs. Create a Firecrawl API key, then add it as a bearer token on the API-key endpoint:

{
  "mcpServers": {
    "firecrawl": {
      "type": "http",
      "url": "https://mcp.firecrawl.dev/v2/mcp",
      "headers": {
        "Authorization": "Bearer <FIRECRAWL_API_KEY>"
      }
    }
  }
}

Replace <FIRECRAWL_API_KEY> with your actual API key. Don't drop the Authorization header, because without it, the Firecrawl MCP connects keyless, which is rate limited and only exposes Search, Scrape, and Parse. Signing in and adding an API key both unlock the full tool surface on your team's plan, including Alexandria's data providers.

Because nothing runs locally, the same config works on macOS, Linux, and Windows.

Method 2: Use the Cursor UI

  1. Click Customize in Cursor's left sidebar, then open the MCPs tab.
  2. Scroll to the bottom and click New MCP Server.
  3. Set the type to Remote HTTPS and the server URL to https://mcp.firecrawl.dev/v2/mcp-oauth, or click Edit JSON and paste one of the configs shared above.
  4. Click Install. If you used the sign-in URL, complete the Firecrawl sign-in when Cursor prompts you.

On older versions of Cursor, the same option lives under Cursor Settings (Cmd+, on Mac, Ctrl+, on Windows) → Tools & MCP → New MCP Server.

Method 3: One-Click Install

Skip the manual config by clicking the button below to configure Cursor automatically:

One-click install button for Firecrawl MCP server in Cursor

Cursor shows the server details before installing. Check that the type is Remote HTTPS and the URL is https://mcp.firecrawl.dev/v2/mcp-oauth:

Cursor's Install MCP server dialog with the Firecrawl hosted server URL

Click Install, then use the server's sign-in option to connect your Firecrawl account and approve access.

Step 2: Verify the connection

After saving your config:

  1. Open Customize → MCPs (or Cursor Settings → Tools & MCP on older versions).
  2. Look for "Firecrawl".
  3. A green dot means it's connected.

Cursor's MCPs list showing Firecrawl connected with 36 tools and 2 resources enabled

Once you're signed in or using an API key, the tool list includes tools like firecrawl_find_tools for discovering Alexandria's data providers. If you only see firecrawl_search, firecrawl_scrape, and firecrawl_parse, the connection is keyless: finish the sign-in or add your API key.

If you see a red dot, check the logs (refresh icon or bottom panel).

Common troubleshooting tips:

  • Server does not connect: Confirm the URL is exactly https://mcp.firecrawl.dev/v2/mcp-oauth (or https://mcp.firecrawl.dev/v2/mcp with your API key header).
  • Sign-in is required or expires: Start authentication again from Cursor's MCP server list.
  • Another Firecrawl connection already exists: If you set up the older local npx firecrawl-mcp server, update it to the hosted URL and sign in so only one Firecrawl server is active.
  • API key issues: Check for the fc- prefix, the Bearer prefix in the header, and no extra spaces.
  • Logs: Use Cmd+Shift+P → "Developer: Show Logs" → "Cursor MCP".

If no Firecrawl tools appear, reload Cursor after saving the MCP configuration.

Step 3: Test the integration

Start a new chat (or open Agent mode with Cmd+I on older versions) and type:

"Use Firecrawl to scrape https://firecrawl.dev and summarize it."

If Cursor asks for approval, click "Allow". The agent calls firecrawl_scrape and summarizes the page.

Cursor agent chat running the firecrawl_scrape tool on firecrawl.dev

Building a competitor price tracker

Let's build a real-world app: a Python script that monitors competitor pricing pages.

1. Create the project:

mkdir pricing-tracker && cd pricing-tracker

2. Prompt the Agent: "I want to build a competitor pricing tracker. Use Firecrawl to scrape the pricing pages of these three SaaS tools and extract the plan names, prices, and key features into a structured format: [Notion, Linear, Todoist pricing URLs]. Then create a Python script that stores this data in a SQLite database and generates a comparison markdown report."

3. Watch it work: The agent recognizes it needs web data and calls firecrawl_scrape on each URL.

Cursor showing firecrawl_scrape approval dialogs for three pricing URLs

After parsing the content (using Firecrawl's built-in JS rendering), it writes your Python script with SQLite storage and a report generator.

Cursor generating the Python pricing tracker script code

The final output is a working script and a generated comparison report.

Final pricing comparison report showing plan details in Cursor

Which tool for which job?

Here is a quick guide on when to use the core web tools:

ToolBest Use Case
firecrawl_scrapeYou know the exact URL and want content fast.
firecrawl_searchYou need to find information but don't know the URL.
firecrawl_extractYou need specific data fields (price, email) in structured JSON.
firecrawl_mapYou want to discover all pages on a site (sitemap).
firecrawl_deep_researchYou need comprehensive analysis across many sources.
firecrawl_agentYou need autonomous browsing/navigation (login, click buttons).
firecrawl_interactYou need to scrape a page then take actions in it (search, click, fill forms). Docs →
firecrawl_batch_scrapeYou have a list of multiple URLs to process in parallel.

Start agentic engineering today

A year after coining "vibe coding," Karpathy proposed "agentic engineering."

Agentic engineering is about giving AI tools real web access and data. Firecrawl MCP in Cursor makes this reality. Your agent can search docs, scrape live sites, and map domains — all without leaving your editor. Beyond coding agents, OpenClaw shows how to give a self-hosted personal agent the same live web capabilities. For the full picture of what MCP servers developers are pairing with Firecrawl today, see the best MCP servers for developers and the top web search MCP servers comparison. If you want to run Firecrawl-powered tasks on a repeating schedule without writing code, see how to schedule recurring web tasks with Claude Desktop or get started with Cursor Automations directly inside the editor.

The fastest way to get started: install Firecrawl directly from the Cursor Marketplace with one click, then open Agent mode and start prompting.

Try Firecrawl MCP for free

Use Firecrawl in your agent

Add Firecrawl to ChatGPT, Claude, Cursor and more in one click.

Frequently Asked Questions

Does Firecrawl MCP work in Cursor's Ask mode?

No. MCP tools only work in Agent mode. Open Composer with Cmd+I (Mac) or Ctrl+I (Windows) and select "Agent".

Do I need to pay for Firecrawl?

Firecrawl offers a free tier with 1,000 credits per month (no card required). This covers ~100 scrapes or 50 searches. Paid plans are available for higher volume.

Can I use Firecrawl MCP with a self-hosted instance?

Yes, but not through the hosted endpoint. Run the open-source MCP server locally and set the FIRECRAWL_API_URL environment variable to your instance (e.g., https://firecrawl.your-domain.com). See Run Firecrawl MCP locally.

I see a red dot next to the server name. What do I do?

Click refresh. If that fails, confirm the server URL is exactly https://mcp.firecrawl.dev/v2/mcp-oauth and start the sign-in again from Cursor's MCP server list. With an API key, use https://mcp.firecrawl.dev/v2/mcp and check the Authorization: Bearer header.

How many MCP servers can I run at once?

Cursor supports ~40 active MCP tools simultaneously.

Will the agent automatically use Firecrawl?

Yes. If you ask to "search the web" or "scrape this site," the agent automatically selects the Firecrawl tool. You can also be explicit: "Use Firecrawl to..."

//
Get started
//
Ready to build?
Start getting web context for free and scale seamlessly as your project expands. No credit card needed.
Are you an AI agent? See setup options