Scraping API PricesAll products / Firecrawl
Firecrawl pricing
How Firecrawl credits vary across scraping, crawling, search, extraction, and browser use, what the keyless tier really allows, what remains open, and who fits.
Every available option was priced in advance from published vendor rules. Changing a control only switches between those checked options; nothing is sent anywhere and no account is needed.
Your job
What Firecrawl counts
What you pay
Successes you want
100,000 pages
Your stated workload. Every vendor on this site is priced against this same number.
- ÷ 0.95
Attempts you are billed for
105,263 requests
Firecrawl bills failed fetches, so a 5% failure rate is billed work.
- × 1
What the meter counts
105,263 credits
One attempt is one billed unit for this workload.
- =
Cash per month
$112.16 /mo
Standard. Cheapest of every eligible purchase option, not the first plan that fits.
Bars share one linear scale across the 3 counted stages, so a multiplier shows as length. Units differ by stage and are labelled. Cash is not on that scale.
100,000 pages divided by 95% success → 105,263 credit → Standard $112.16/mo = $1.12 per 1,000 pages kept
The arithmetic, line by line
- pages: 1 credit per pagesource
- request.js_render: no multiplier published, treated as included
- billed on 105,263 attempts, not 100,000 successes, because this vendor charges for failed requestssource
- consumes 105,263.16 credit per month
- cheapest way to buy that: Standard plus 5,263 over the allowance at $0.0025 each = $112.16 per monthsource
Cheaper for this job
Same 100,000 JavaScript-rendered pages a month at a 95% success rate, ranked by what you would actually be billed. Unknown totals never enter the numeric order. Showing 6 of 9; see the rest.
Every Firecrawl plan
| Plan | Monthly | If billed yearly | Prepaid pack | Included | Simultaneous sessions |
|---|---|---|---|---|---|
| Hobby | $19.00 | $16.00 | n/a | 5,000 credit | 5 |
| Standard your job | $99.00 | $83.00 | n/a | 100,000 credit | 25 |
| Growth | $399.00 | $333.00 | n/a | 500,000 credit | 50 |
| Scale | $749.00 | $599.00 | n/a | 1,000,000 credit | 100 |
What consumes a unit
This is where a bill actually grows. One operation is not one billed unit, and the multiplier is published here rather than on the pricing page.
| Operation | Costs | Per | Source |
|---|---|---|---|
| request.basic | 1 credit | page | source |
| crawl.page | 1 credit | page | source |
| search.query | 2 credit | 10 result | source |
| browser.session_time | 2 credit | browser_minute | source |
| request.ai_extraction | 5 credit | page | source |
| request.premium_proxy | 1 credit | request | source |
| request.pdf_parse | 1 credit | page | source |
What Firecrawl does not publish
- The fixed credit cost of Agent is not published.
- The exact token direction and payload accounting for Extract are not published in the selected billing guide.
- The per-IP daily request and credit caps that bound keyless use are not published, so the ceiling on no-signup usage cannot be stated as a number.
- The schema has one meters_combine field, but Firecrawl mixes replacement pricing, additive features, and separately billed browser time.
What Firecrawl is
Firecrawl is a web data platform with scrape, crawl, map, search, monitor, structured extraction, agent, and browser-interaction operations. It is aimed at developers building retrieval, research, enrichment, and agent workflows that need web content in application-ready formats. Olostep, Hyperbrowser, and Apify are the nearest named comparisons. Firecrawl differs from them by centering one API family on turning sites into content and context while also exposing search, an agent, and a browser sandbox. Sources: Firecrawl pricing, Firecrawl browser documentation, Olostep, Hyperbrowser pricing, and Apify pricing, retrieved 2026-09-01.
How billing works
The shared accounting unit is a credit, but the workload unit changes by feature. Basic scrape, crawl, map, and monitor usage follows pages; search follows returned-result groups; browser interaction follows elapsed browser minutes; structured extraction can consume more credits; and Agent uses dynamic pricing. Some browser-interaction paths also bill the initial scrape separately. Firecrawl does not charge Firecrawl-side timeouts or server errors, but it does charge when it fetches a page and the target site returns an error response. The common buying mistake is to equate one page with one credit across every endpoint. Sources: Firecrawl pricing, browser documentation, Interact documentation, and extraction guide, retrieved 2026-09-01.
Running out of credits mid-month is not a wall. Auto-reload buys more in $5 batches, and the batch size is what sets the marginal rate: $5 buys 1,000 credits on Hobby, 2,000 on Standard, 2,500 on Growth, and 5,000 on Scale. That works out at $0.005, $0.0025, $0.002, and $0.001 per credit respectively, and it is why the cheapest way to run a large job is often a mid-tier plan plus overage rather than the next plan up. A monthly auto-reload limit caps the spend, and setting it to zero turns auto-reload off. Credits do not roll over on self-serve plans; Firecrawl lists rollover as a Scale and Enterprise feature. Source: Firecrawl pricing, retrieved 2026-09-01.
The keyless tier, and what actually bounds it
Since 2026-08-28 Firecrawl’s free allowance can be used with no signup and no API key at all. Search, Scrape, and Parse accept unauthenticated calls, through the hosted MCP endpoint at https://mcp.firecrawl.dev/v2/mcp, through the CLI, and through plain REST with no Authorization header. The allowance is the same one a free account gets: 1,000 credits every month, with no card. Sources: keyless launch announcement and Firecrawl pricing, retrieved 2026-09-01.
Keyless is not unlimited, and the limit is the part worth reading twice. Keyless use is capped per IP address, per day, by two independent ceilings: a maximum request count and a maximum credit count. Firecrawl publishes neither number. Heavier operations consume more credits, so the credit ceiling can be reached well before the request ceiling, and exceeding either returns HTTP 429. A shared office address, a CI runner, or a pool of cloud functions all present as one IP, so the practical keyless budget is not one allowance per developer. Source: Firecrawl rate limits, retrieved 2026-09-01.
Separately from the keyless caps, every account has per-minute rate limits, enforced per team across all of its API keys. Free allows 10 scrape, 10 map, and 10 search requests a minute, and 2 crawl and 2 agent requests a minute. Standard allows 500 scrape, 500 map, and 500 search, and 100 crawl and 100 agent. Extract shares the agent limit and batch scrape shares the crawl limit. Source: Firecrawl rate limits, retrieved 2026-09-01.
What Firecrawl does not publish
- A fixed credit cost for Agent, which the pricing page labels dynamic. Source reviewed: Firecrawl pricing, retrieved 2026-09-01.
- A complete public conversion for structured extraction that separates input, output, and other token accounting. Source reviewed: Firecrawl extraction guide, retrieved 2026-09-01.
- The two per-IP daily caps that bound keyless use. Both are described and neither is numbered, so the ceiling on no-signup usage cannot be planned against. Source reviewed: Firecrawl rate limits, retrieved 2026-09-01.
- One combination rule covering base scrapes, enhanced mode, browser interaction, extraction, and Agent. Sources reviewed: Firecrawl pricing, Interact documentation, and enhanced mode documentation, retrieved 2026-09-01.
What changed since the last reading
Two of the numbers on this page moved in the buyer’s favour. Enhanced mode, which Firecrawl previously documented as stealth mode, now costs the same as a basic request, 1 credit, where it used to cost 5. And the overage rate above a plan’s included credits, previously recorded here as unpublished, is now readable from the $5 auto-reload batch sizes. Two numbers moved the other way: concurrency on Standard is 25 and on Growth is 50, lower than the 50 and 100 recorded here before. Sources: enhanced mode documentation and Firecrawl pricing in markdown, retrieved 2026-09-01.
Who it suits
Firecrawl suits teams that want scrape, crawl, search, and browser interaction under one developer API and can measure the exact endpoint mix. The keyless tier makes it unusually good for the first hour of a project, when the question is whether a site can be read at all and a signup form is the main thing in the way. It is less suitable for procurement that needs one fixed cost per successful page regardless of feature, for an Agent-heavy budget that requires a published fixed run rate, or for any plan that leans on keyless access at volume, because the caps that bound it are per IP and unpublished. Sources: Firecrawl pricing, rate limits, and browser documentation, retrieved 2026-09-01.
Answer
Firecrawl costs $112.16 for 100,000 pages under stated assumptions. It is not ranked until every billing rule is known.
Recorded 2026-09-01 from the vendor's own pricing page.This published price is inside the 30-day comparison window. Its date stays visible because it was not confirmed by the latest automated check. Confirm it on the vendor's pricing page before buying.