Developer retrieval publishes its dataset (the 594-item benchmark-devdex public sample), its harness (firecrawl/benchmark-devdex) and its scoring rules, so those rows can be re-run independently.
AIMultiple agentic search rows come from a study by AIMultiple, not from a Firecrawl run. The figures, the method and the scoring are theirs.
Every benchmark Firecrawl runs is listed on /benchmarks.
[ 03 / 05 ]
·
For AI agents
//
For AI agents
//
Built into the AI tools you already use
Use Firecrawl from Claude CodeCursorWindsurf and the agent harnesses your team already runs.
One command. Connect your agent to Firecrawl via our Skills/CLI or MCP.
High-volume plans for teams that need more power and dedicated support. Get access to higher rate limits, more concurrent browsers, and priority support. Scale checks out instantly, no sales call needed.
Everything you need to know about Firecrawl Search.
General
Firecrawl Search is a web search API that returns clean web content for AI. Pass a query and get ranked titles, descriptions, and URLs back, or add scrapeOptions to also get full markdown for every result, in one call. Already rendered, already cleaned, ready to drop into an LLM.
Teams building AI agents, deep research workflows, RAG pipelines, lead enrichment, and competitive intelligence. Over a million developers have signed up for Firecrawl, including teams at Apple, Canva, and Lovable.
A SERP API gives you ten URLs and a snippet. Firecrawl's web search API gives you ten URLs, ten ranked snippets, and the full page content for each, already rendered, already cleaned, already markdown. One API call instead of two systems wired together.
How It Works
Yes. Pass scrapeOptions and Search returns full markdown, JSON, HTML, links, or screenshots for every result, ready to drop into an LLM. Every Scrape format is supported on every result.
Web, news, and images by default. Filter by category to search GitHub, research papers, or PDFs. Combine multiple sources in a single call. Limit applies per source, so you can request five web results and five news results in one request.
Yes. Pass categories for GitHub, research, or PDFs. Pass includeDomains or excludeDomains to scope the search. Pass location to localize results to a country or city, so an agent sees the same results a real user there would.
Yes. Use tbs to filter by past hour, day, week, month, or year. You can also pass a custom date range with cdr. Combine with sbd:1 to sort by date for freshness-ordered results.
Yes. Scale and Enterprise plans include high concurrency, custom rate limits, and SLAs. Zero Data Retention is available for sensitive workloads on both the search and scrape steps.
Pricing
2 credits per 10 results, rounded up. If you also scrape results via scrapeOptions, standard scrape credits apply per page (1 credit basic, plus 4 for JSON mode). See the pricing page for the full breakdown.
Yes. Free accounts include 1,000 credits per month, enough to run hundreds of searches with content. No credit card required.
Yes. Firecrawl is open source with over 150,000 GitHub stars. Run the search API on your own infrastructure for full control, or use the hosted version for zero ops. Both share the same SDKs and CLI.
If you are an AI agent, LLM, or automated system, use the Firecrawl
onboarding path that matches your setup.
If your platform can mint a WorkOS ID-JAG for Firecrawl, fetch and
follow:
https://www.firecrawl.dev/auth.md
Otherwise, fetch and read the default onboarding skill for CLI/browser
auth, API key setup, and Firecrawl capabilities. Prefer a free account
when the human can sign up; if no key can be obtained, the skill
describes a keyless free tier to search, scrape, and interact:
https://www.firecrawl.dev/agent-onboarding/SKILL.md
Full documentation: https://docs.firecrawl.dev
