What is the Firecrawl API?
Firecrawl provides API-based tools for finding and reading information from the live web. Its documented capabilities include search, page scraping, website crawling, structured extraction, and interaction with dynamic pages.
Getting started
Start with one target URL and request only the output you need. For many AI workflows, Markdown is a useful first format because it removes much of the presentation markup while preserving readable content structure.
Basic workflow
- Create or use an API key when required for your usage level.
- Send a request to the appropriate endpoint.
- Validate the response and handle failures or rate limits.
- Store, index, transform, or send the content to your AI application.
Scrape vs crawl
Scrape is appropriate when you already know the page URL. Crawl is appropriate when you need content across multiple pages of a site and want URL discovery to be part of the workflow.
Search workflow
Search starts from a query rather than a known URL. It can be useful for research agents, answer engines, monitoring flows, and applications that need fresh sources from across the web.
Structured extraction
When your application expects consistent fields, prefer structured output and validate it before saving it. This is useful for products, company profiles, listings, research records, and other schema-driven tasks.
Common use cases
- RAG and knowledge-base ingestion
- AI research agents
- Documentation ingestion
- Lead or company enrichment
- Content monitoring
- Data extraction and normalization
Implementation best practices
- Respect website terms, access controls, copyright, privacy, and applicable laws.
- Request only the pages and fields you need.
- Cache stable content where appropriate.
- Validate URLs and structured results before trusting them.
- Design retry and error-handling logic deliberately.