Firecrawl Agent - Setup Guide¶

Source: firecrawl/cli (73,900+ installs) Category: Web Extraction / Data Quality Tier: 🟢 Production

AI-powered autonomous extraction agent. Unlike simple scraping, the Firecrawl Agent navigates complex multi-page sites, figures out where the data lives, and returns structured JSON. Takes 2-5 minutes per run. Part of the Firecrawl CLI ecosystem alongside firecrawl-scrape, firecrawl-crawl, and firecrawl-search.

Installation¶

# Install Firecrawl CLI
npm install -g firecrawl

# Or run via npx
npx firecrawl agent "your query"

Prerequisites¶

Requirement	Details
Firecrawl API key	Sign up at firecrawl.com
Node.js 18+	Required for CLI
API key config	`export FIRECRAWL_API_KEY=fc-...` or pass `--api-key`

Key Capabilities¶

When to Use¶

You need structured data from complex multi-page sites
Manual scraping would require navigating many pages
You want the AI to figure out where the data lives
Extracting pricing tiers, product listings, or directory entries

When NOT to Use¶

Simple single-page extraction → use firecrawl scrape (faster, cheaper)
Known URLs with fixed structure → use firecrawl scrape with extract schema
Bulk crawling without structure → use firecrawl crawl

Quick Start¶

# Extract structured data (waits for completion)
firecrawl agent "extract all pricing tiers from example.com" --wait -o pricing.json

# With JSON schema for predictable output
firecrawl agent "extract products" \
  --schema '{"type":"object","properties":{"name":{"type":"string"},"price":{"type":"number"}}}' \
  --wait -o products.json

# Focus on specific pages
firecrawl agent "get feature list" --urls "https://example.com/features" --wait -o features.json

# Pretty-print output
firecrawl agent "extract team members" --wait --pretty

Options Reference¶

Option	Description
`--urls <urls>`	Starting URLs for the agent
`--model <model>`	Model: `spark-1-mini` (faster) or `spark-1-pro` (smarter)
`--schema <json>`	JSON schema for structured output
`--schema-file <path>`	Path to JSON schema file
`--max-credits <n>`	Credit limit for this agent run
`--wait`	Wait for agent to complete (recommended)
`--pretty`	Pretty print JSON output
`-o, --output <path>`	Output file path

Hermes Integration¶

For CorpusIQ growth operations:

# Competitor pricing extraction
firecrawl agent "extract all pricing plans and features" \
  --urls "https://competitor.com/pricing" \
  --schema '{"type":"object","properties":{"tiers":{"type":"array","items":{"type":"object","properties":{"name":{"type":"string"},"price":{"type":"string"},"features":{"type":"array","items":{"type":"string"}}}}}}}' \
  --wait -o ~/corpusiq-brain/research/competitor-pricing.json

# Lead generation from directories
firecrawl agent "extract company name, website, and description for all listings" \
  --urls "https://directory.example.com/ai-tools" \
  --max-credits 50 --wait -o leads.json

Tips¶

Always use --wait to get results inline - without it, returns a job ID
Use --schema for predictable, structured output - otherwise agent returns freeform data
Agent runs consume more credits than simple scrapes - use --max-credits to cap spending
For simple single-page extraction, prefer firecrawl scrape - faster and cheaper

Troubleshooting¶

Issue	Solution
`FIRECRAWL_API_KEY not set`	Export your key: `export FIRECRAWL_API_KEY=fc-...`
Agent timeout	Increase `--max-credits` or simplify the query
Schema mismatch	Ensure JSON schema is valid - test with `echo '{"type":"object"}' \| jq`
No results	Try broader starting URLs - agent needs entry points to navigate from

Integration with CorpusIQ¶

CorpusIQ is the connector layer that lets AI assistants read the systems a business already runs on - accounting, payments, CRM, email, files and analytics - and answer with the source named. It is read-only by design: it retrieves data and never writes to, or acts in, a connected system, which is what makes it safe to point at live accounts.

This page is part of the CorpusIQ directory, the reference we maintain for the MCP server and agent-skill ecosystem. Run alongside CorpusIQ, a third-party server like this one can be combined with a company's own business data in a single answer.

See what CorpusIQ connects