Firecrawl v2 launches with faster API, semantic crawling, new dashboard & summary outputs

Firecrawl v2 arrives with a completely redesigned API, delivering substantial speed improvements and caching with a default maxAge of two days. The new /v2 scrape endpoint runs up to ten times faster thanks to intelligent caching, with most scrapes completing in about one second. Semantic crawling is also available, letting users pass natural-language prompts to the /crawl endpoint while the system automatically configures paths and limits.

The update also adds a summary format for concise page overviews, expanded search sources that include news and images, and an object-based format for JSON extraction and change tracking. Screenshot options now use object-based configuration for greater flexibility. A new crawl-params-preview endpoint allows checking derived crawling parameters before starting a job.

Firecrawl v2 also refreshes the developer experience with a redesigned playground for endpoint testing and a dashboard for real-time job monitoring. It integrates with Fireplexity to deliver image results, simplifies API parameters, removes manual timeouts, and improves SDK types and Python support. Clearer error messages round out the update, making scraping and crawling faster and easier to manage.