Firecrawl pricing and review

Reachkit, which publishes this page, makes cold email software with a built-in lead database, so it overlaps with Firecrawl only where Firecrawl now resells contact data. Every price below was checked on October 15, 2026 against Firecrawl’s pricing page, billing docs and terms.

Firecrawl pricing starts free with 1,000 credits a month, and the paid plans run from $19 to $749 a month, or $16 to $599 billed yearly. One credit scrapes one plain page. Figures from firecrawl.dev/pricing and the billing docs, checked October 15, 2026.

Firecrawl’s page prints a monthly price only for Hobby. The Standard, Growth and Scale monthly prices are our math from the yearly savings it shows, and match third-party price guides from July and August 2026. Every plan but Enterprise is bought without a sales call.

What a credit buys. Scrape and crawl cost 1 credit a page, and search costs 2 credits per 10 results. Options stack on top: JSON extraction by an AI model adds 4, so an extracted page costs 5, and PDF parsing adds 1 per PDF page. X (Twitter) pages cost 30 credits, and LinkedIn data is priced by the provider behind it, from 5 to 40. A scrape that returns nothing is free, but a returned 403 or 404 page costs 1.

Running out. Paid plans can top up automatically in $5 steps, from $5 per 1,000 credits on Hobby down to $1 on Scale. You set a monthly limit in Billing, and a blank limit means no cap. Hobby, Standard and Growth credits don’t roll over; Scale carries them one month (yearly Scale only, per the billing docs). Upgrades are charged in full, not prorated, and the terms rule out refunds unless Firecrawl stops offering the service.

What a typical setup costs. Extracting fields from 2,000 company websites a month, one page each, takes 10,000 credits: about $44 on monthly Hobby ($19 plus $25 of top-ups). At 20,000 plain pages, Hobby with top-ups reaches $94, close to Standard’s $99 for 100,000 credits, so past that point Standard is the better deal.

Firecrawl fetches web pages for you, renders their JavaScript, handles proxies and hands back the content in a format a program or an AI model can read. You call it with a URL, a whole site or a search query and get markdown, HTML, structured JSON or a screenshot.

The main endpoints are Scrape (one page), Crawl (follow links across a site), Map (list a site’s URLs) and Search (search the web and scrape the results). Newer ones add a live browser, page monitoring and an Agent, still in preview, that gathers data from a prompt.

Since September 2026 Firecrawl also sells Alexandria, a library of data providers for AI agents, which added people and company enrichment from Apollo, FullEnrich and Data Legion in October 2026.

The core is open source under AGPL-3.0, and the hosted version adds features such as its Fire-engine. Firecrawl also runs an official MCP server, a standard way for AI assistants like Claude and ChatGPT to use other apps, with SDKs for Python, Node.js, Go, Rust, Java and more.

Strengths

• A free tier that renews. 1,000 credits every month with no card, plus a keyless tier with daily limits for search, scrape and parse (pricing, MCP docs).
• Output ready for AI. Clean markdown and good handling of JavaScript-heavy pages are the most common praise, on Product Hunt and in Gumloop’s review.
• Open source. The AGPL-3.0 core has 191.7k stars on GitHub and can be self-hosted, though the cloud version has extra features.
• No charge for empty results. Empty scrapes aren’t charged, and the enhanced proxy has cost the same 1 credit as a basic request since August 17, 2026 (changelog).
• Built for agents. A hosted MCP server with sign-in or API-key access.

Drawbacks

• Credits per page add up. JSON extraction makes a page cost 5 credits, other options add 4 each, and a 403 or 404 page still costs 1 (billing docs).
• No rollover below Scale. A spiky month and a quiet month are billed the same.
• Strict money terms. No refunds “for any reason” (terms), and top-ups run uncapped if the limit is left blank.

Firecrawl has few written reviews for its size. G2 shows 4.6 from just 6 reviews, read from search results, and Capterra lists it with no reviews yet, also per search results. There’s no Trustpilot page for Firecrawl, checked October 15, 2026. Product Hunt shows 5.0 from 17 reviews, 16 of them from founders.

The praise is consistent: clean output that saves parsing raw HTML, and good handling of dynamic pages. Much of the criticism comes from review sites run by competing scrapers or other software vendors. The recurring points are credits that don’t roll over, add-ons that make costs hard to predict, and mixed results on heavily protected sites such as LinkedIn and Amazon.

Firecrawl fits developers, AI products and GTM engineers who need web page content inside a workflow and would rather pay per page than run scrapers and proxies. The free tier makes testing cheap.

Look elsewhere if nobody on your team writes code, or if your volume swings so much that credits go unused. For cold email, a scraped homepage tells you what a company does, not who to email. AI earns its keep in list building, so I’d use a tool like this to qualify accounts (do they sell to agencies, do they hire in sales) rather than to write the emails. You still need a named contact and a verified email. More in lead scraping for cold email.

Crawl4AI, Apify, ScrapeGraphAI and Spider are scraping tools people compare with Firecrawl; we haven’t checked their prices. For turning companies into contacts:

• Clay: enrichment tables that run many data providers on each row, with a free plan and Launch at $185 a month. See the Clay profile.
• Apollo: a contact database with sequences built in, a free plan and Basic at $65 a seat a month. See the Apollo profile.
• PhantomBuster: ready-made cloud automations that scrape LinkedIn and other sites, from $69 a month. See the PhantomBuster profile.
• Hunter: finds and checks emails for a company domain, with a free plan and Starter at $49 a month. See the Hunter profile.
• Reachkit (our product): cold email software whose Lead Finder (500M+ people, 70M+ companies) supplies the named contact and verified email a scraped page lacks, at 1 credit per lead. It doesn’t scrape websites; import your page research by CSV as merge tags. From $39/month.