Firecrawl API Reference and Usage Guide (2026)

title: "Firecrawl API Reference and Usage Guide (2026)" author: Zachary Proser date: 2026-4-18 description: A practical guide to the Firecrawl API — authentication, endpoints, rate limits, pricing, and how to integrate Firecrawl into your web scraping and data extraction workflows. image: https://zackproser.b-cdn.net/images/firecrawl-hero.webp tags: [ai, web-scraping, firecrawl, api, developers] hiddenFromIndex: true

Firecrawl provides a clean API for scraping websites at scale, with particular strength in handling JavaScript-rendered pages and extracting structured data. The API design is straightforward: authenticate, call an endpoint, get back clean Markdown or structured JSON.

Here is what you actually need to know to use it in production.

Authentication

You get your API key from the Firecrawl dashboard. Pass it as a Bearer token in the Authorization header:

Authorization: Bearer fc-your-api-key-here

The free tier gives you 500 credits. Each API call consumes credits based on the number of pages scraped and whether you use the async batch endpoint.

Core Endpoints

POST /v1/scrape — scrape a single URL and get back structured data. You can specify which fields to extract using a schema.

import requests

response = requests.post(
    "https://api.firecrawl.dev/v1/scrape",
    headers={"Authorization": "Bearer fc-your-key"},
    json={
        "url": "https://example.com/pricing",
        "extractorConfig": {
            "mode": "llm-extraction"
        },
        "pageRanges": ["1-5"]
    }
)
data = response.json()
print(data["data"]["content"])  # Markdown content

POST /v1/batch/scrape — submit multiple URLs for asynchronous scraping. Returns a job ID you poll until completion. This is the right choice for anything more than 5-10 URLs.

POST /v1/crawl — crawl an entire website starting from a seed URL, following internal links within the specified limits.

POST /v1/extract — given a URL and a prompt, extract structured data matching your schema. This is the most powerful endpoint for building datasets.

response = requests.post(
    "https://api.firecrawl.dev/v1/extract",
    headers={"Authorization": "Bearer fc-your-key"},
    json={
        "url": "https://news.ycombinator.com",
        "prompt": "Extract all article titles, scores, and author names"
    }
)
articles = response.json()["data"]["entities"]

Rate Limits and Pricing

The free tier: 500 credits/month. A single page scrape consumes 1 credit. Batch scraping uses 0.5 credits per page.

Paid plans start at $15/month for 5,000 credits. The growth plan at $49/month gives you 25,000 credits and priority rate limits.

Rate limits on the free tier: 20 requests/minute. Paid tiers get 60-200 requests/minute depending on your plan.

For production scraping jobs, batch endpoints are significantly more efficient — you submit once and poll, rather than hammering the synchronous endpoint.

Handling JavaScript-Rendered Pages

Firecrawl uses a headless browser stack under the hood. Pages that require JavaScript execution are handled automatically. The pageRanges parameter lets you specify how many pages to scroll through on paginated content.

For single-page applications that lazy-load content, you can increase the waitFor timeout in the request options.

Common Issues

Incomplete content on SPA pages — If a page loads content via JavaScript after the initial render, you may need to increase the waitFor timeout or use the batch scrape endpoint which gives pages more time to fully render.

CAPTCHA and bot detection — Firecrawl handles some but not all anti-bot measures. For sites with aggressive protection, you may need to combine with proxy rotation.

Large websites — Crawling a site with thousands of pages quickly consumes credits. Use page range limits and be selective about which sections you crawl.

When to Use Firecrawl vs Crawl4AI

Crawl4AI is better for developers who want full control over the crawling process and are comfortable with a more hands-on setup. Firecrawl is the right choice when you want a managed service with reliable uptime, structured data extraction without writing custom parsers, and batch processing at scale.

For RAG pipelines specifically, Firecrawl's extract endpoint is the fastest way to get clean, structured text from a list of URLs.