Cost & Pricing

Updated Feb 2026

Credit costs for all Firecrawl operations.

Credit Usage

Operation	Credits	Notes
Scrape (single URL)	1	Per page
Batch Scrape	1	Per page
Search	2	Per 10 results
Search + Scrape	2 + scrape costs	Per 10 results + per page scraped
Map	1	Per call (any number of URLs)
Crawl	1	Per page crawled
Browser Sandbox	2	Per minute of session time
Agent (Spark-1 Mini)	Dynamic	~60% cheaper than Pro
Agent (Spark-1 Pro)	Dynamic	Most runs = a few hundred credits
Agent (Spark-1 Fast/CSV)	10	Per cell (parallel batch)

Add-On Costs

Add-On	Credits	When Used
JSON Mode (scrape)	+4	Extracting structured data via schema
Enhanced Proxy	+4	Complex websites / anti-bot protection
PDF Parsing	+1 per PDF page	Processing PDF documents
Change Tracking (basic + git-diff)	Free	No extra cost over normal scrape
Change Tracking (JSON mode)	+5	Field-level comparison with schema
Screenshot	Free	Included with scrape
Branding extraction	Free	Included with scrape

Free Tier

Feature	Free Allowance
Agent runs	5 free daily runs (all users)
Browser usage	5 hours of free usage
API credits	1,000 credits (Starter plan)

Cost Examples

Basic Website Scrape

• 1 page = 1 credit

JSON Extraction

• 1 page with schema = 5 credits (1 + 4)

Website Crawl (100 pages)

• 100 pages = 100 credits

Search & Scrape (10 results)

• Search + scrape 10 pages = 12 credits (2 + 10)

Browser Session (5 minutes)

• 5 minutes = 10 credits

Agent Research

• Simple query (Spark-1 Mini) = ~100-300 credits
• Complex analysis (Spark-1 Pro) = ~200-500 credits

Change Monitoring (50 pages daily)

• 50 scrapes/day with changeTracking = 50 credits/day
• With JSON diff mode = 300 credits/day (50 x 6)

Cost Optimization

DO

• Use caching (maxAge) — up to 5x faster, reuses recent results
• Set crawl limits — always set limit to control page count
• Use basic proxy — don't use enhanced unless needed
• Batch operations — batch scrape for efficiency
• Use spark-1-mini — 60% cheaper, good enough for most Agent tasks
• Set maxCredits on Agent — caps runaway costs
• Use change tracking tags — avoid redundant tracking histories

DON'T

• Crawl without setting limit (default is 10,000 pages!)
• Use enhanced proxy for simple sites
• Re-scrape recently scraped content (maxAge covers this)
• Extract JSON when markdown is sufficient
• Set maxAge=0 unnecessarily (slower, more likely to fail)
• Leave browser sessions open (credits tick per minute)

Monitoring Usage

Dashboard

• Login: firecrawl.dev/app
• Navigate: Logs > Activity
• View: Credits used, remaining balance

API

# Current credit usage
curl -X GET 'https://api.firecrawl.dev/v2/account/credit-usage' \
  -H 'Authorization: Bearer fc-YOUR-API-KEY'

# Historical usage
curl -X GET 'https://api.firecrawl.dev/v2/account/credit-usage/historical' \
  -H 'Authorization: Bearer fc-YOUR-API-KEY'

MCP Credit Alerts

Set thresholds in your MCP config:

FIRECRAWL_CREDIT_WARNING_THRESHOLD=1000   # Warning at 1000 credits
FIRECRAWL_CREDIT_CRITICAL_THRESHOLD=100   # Critical at 100 credits

---

Full Pricing