Skip to Content

Documentation that starts with working code

Keep docs flatter and quieter than marketing pages so code examples and operational steps carry the proof.

Built for developers

The code surface should look like a real operator workflow, not placeholder SDK marketing.

main.py
1 # pip install firecrawl-py
2 from firecrawl import FirecrawlApp
3
4 app = FirecrawlApp(api_key="fc-***")
5
6 # Search, scrape, then hand clean markdown to your agent
7 result = app.scrape_url("https://firecrawl.dev")
8 print(result["markdown"])

Everything your agents need to read the web

Keep the visual language technical and proof-driven: capability first, output second, trust third.

Search
Find the right page first
Query the live web and pull back the URLs, snippets, and signals your agent actually needs.
Scrape
Return markdown or structured data
Convert noisy pages into clean payloads your workflow can reason about immediately.
Interact
Handle live browser actions
Step through dynamic flows when a static fetch is not enough, without leaving the same product surface.
START SCRAPING TODAY

Start with one API. Scale when it proves itself.

Keep the CTA direct and operator-friendly: start for free, check the docs, then move into higher-throughput plans.