Firecrawl for AI agents

Firecrawl crawls websites and extracts structured content at scale, used to gather data for search, research, and indexing by teams building AI and data products. Connect it once through Arc0, and your agent, or Claude, ChatGPT and Cursor, can use it through one MCP endpoint, limited to what each user approved.

Connect Firecrawl See the 30 actions ↓

Web data and scrapingAPI keyMCP + RESTfirecrawl.dev

AUDIT LOG · FIRECRAWLPOLICY: acme-support

09:41:07 · claude · u_8f2read

firecrawl.search

Search✓ allowed · 212ms

09:41:08 · claude · u_8f2read

firecrawl.crawl

Start a web crawl✓ allowed · 164ms

09:41:09 · claude · u_8f2write

firecrawl.queue_get

Get team queue status✓ approved · approved by user

09:41:12 · claude · u_8f2destructive

firecrawl.agent_cancel

Cancel an agent job✕ blocked · policy: deny

EVERY FIRECRAWL CALL, ON THE RECORD

01 · USE CASES

What agents do in Firecrawl.

Scrape a single URL

Scrape a URL to pull its clean content for a summary or research task.

Check a crawl job's status

Get a crawl job's status to see how many pages have been processed so far.

Approve before a large crawl

Confirm the scope of a site-wide crawl before starting it, since it consumes team credits at scale.

02 · ACTIONS

30 Firecrawl actions, graded by risk.

Every Firecrawl action is tagged read, write or destructive, so one policy covers the whole app and new actions inherit the right default.

read

Look things up. Allowed by default.

firecrawl.get_agent_status
Get agent job status
firecrawl.get_deep_research_status
Get deep research status
firecrawl.get_the_status_of_a_crawl_job
Get the status of a crawl job
firecrawl.crawl
Start a web crawl
firecrawl.scrape
Scrape URL
firecrawl.search
Search
firecrawl.extract
Extract structured data
firecrawl.crawl_v2
Start a web crawl (v2) [NEW]
firecrawl.crawl_get
Get crawl job status
firecrawl.extract_get
Get extract job status
firecrawl.crawl_get_errors
Get errors from a crawl job
firecrawl.crawl_list_active
Get all active crawl jobs
firecrawl.crawl_params_preview
Preview crawl parameters

write

Create and change things. Allow, or ask the user first.

firecrawl.queue_get
Get team queue status
firecrawl.start_agent
Start an agent job
firecrawl.batch_scrape
Batch scrape multiple URLs
firecrawl.llms_txt_get
Get LLMs.txt generation job status
firecrawl.deep_research
Perform deep research
firecrawl.token_usage_get
Get team token usage
firecrawl.batch_scrape_get
Get batch scrape status
firecrawl.credit_usage_get
Get team credit usage
firecrawl.llms_txt_generate
Generate LLMs.txt for a website
firecrawl.batch_scrape_get_errors
Get errors from batch scrape job
firecrawl.token_usage_get_historical
Get historical team token usage
firecrawl.credit_usage_get_historical
Get historical team credit usage
firecrawl.map_multiple_urls_based_on_options
Map multiple URLs

destructive

Delete, cancel or archive. Ask first, or deny outright.

firecrawl.agent_cancel
Cancel an agent job
firecrawl.crawl_cancel
Cancel a crawl job
firecrawl.crawl_delete
Cancel a crawl job
firecrawl.batch_scrape_cancel
Cancel a batch scrape job

03 · HOW IT WORKS

Firecrawl in three steps.

01Your users connect FirecrawlThey add their Firecrawl api key on Arc0 Connect, under your brand. It goes straight into the vault.
02You set the rulesReads run, writes like “get team queue status” can wait for the user, and “cancel an agent job” can be denied outright.
03Any agent can actYour agent calls Firecrawl through the Arc0 SDK or MCP, and so can Claude, ChatGPT and Cursor. Every call lands on the audit log.

POLICY.TS

await arc0.policies.set('firecrawl', {
  read: 'allow',
  write: 'ask',        // queue_get
  destructive: 'deny',  // agent_cancel
})

# Claude Code: the same connection, one URL
$ claude mcp add --transport http arc0 \
    https://mcp.arc0.ai/u/u_8f2

04 · AUTH AND DATA

How Firecrawl connects.

Users add their Firecrawl api key on Arc0 Connect. It is encrypted in the vault, never shown to the model, and each user can rotate or revoke it at any time.

The same Firecrawl connection serves your agent over MCP and your own backend over REST and the proxy, so a user connects once. How Arc0 handles credentials →

AUTH: API key
CREDENTIALS: Per-tenant encrypted vault
MODEL SEES: Results only, never credentials
AUDIT LOG: Every call, on every plan

05 · WORKS WITH

Use Firecrawl from any agent.

Claude

ChatGPT

Cursor

Codex

VS Code

OpenAI Agents SDK

Claude Agent SDK

Vercel AI SDK

Mastra

LangGraph

06 · RELATED

More web data and scraping apps.

Just One API288 actions

Apify183 actions

OpenWeb Ninja164 actions

ALL WEB DATA AND SCRAPING APPS →

07 · FAQ

Firecrawl and Arc0, answered.

Q01

Can I use Firecrawl with Claude, ChatGPT or Cursor?

Yes. Connect Firecrawl to Arc0 once, then add your Arc0 MCP URL to Claude, ChatGPT, Cursor, Claude Code or any other remote-MCP client. Each assistant only gets the Firecrawl actions you allow.

Q02

How do users connect Firecrawl?

Users add their Firecrawl api key on Arc0 Connect. It is encrypted in the vault, never shown to the model, and each user can rotate or revoke it at any time.

Q03

Which Firecrawl actions can my agent take?

30 in total: 13 read, 13 write and 4 destructive, such as “get team queue status”. Your policies decide which of them each agent may call.

Q04

Can I stop my agent from deleting things in Firecrawl?

Yes. Actions like “cancel an agent job” are graded destructive. Set destructive actions to deny, or to ask so the user approves each one, and blocked calls still show up on the audit log.

Q05

Can my own backend call Firecrawl too?

Yes. The same Firecrawl connection is available over REST and through the Arc0 proxy, so your product and your agent share one connection per user.

Get started

Plug Firecrawl into your agent.

Your users connect Firecrawl once, under your brand. Your agent gets 30 actions behind your policies, with every call on the record.

Connect Firecrawl Browse all apps →

Free to build · MCP + REST · Audit log on every plan