Firecrawl Pricing: Free, Hobby, Standard & Scale
Start on Free if you stay under 1,000 basic scrape credits a month. Buy Hobby ($19/month, or $16/month billed annually) when you need JSON extraction, a little more volume, or 5 concurrent browsers. Buy Standard ($99/month, or $83/month billed annually) once production Markdown or JSON crosses Hobby’s 5,000 credits — 100,000 credits is the step most teams actually land on.
- Official hosted plans, checked November 6, 2026: Free $0 / 1,000 credits / 2 browsers; Hobby $19 or $16 annual display ($190/year) / 5,000 / 5; Standard $99 or $83 ($990/year) / 100,000 / 25; Growth $399 or $333 ($3,990/year) / 500,000 / 50; Scale $749 or $599 ($7,190/year) / 1,000,000 / 100; Enterprise custom.
- Credits are not 1:1 with pages. Basic scrape or crawl is 1 credit per page. A JSON scrape is 5. Search is 2 credits per 10 results, rounded up. Interact is billed per browser minute.
- Worth-it vs which-plan: this page picks a plan. Whether Firecrawl is the right product at all lives on our Firecrawl review.
By Unfair Advantage · Updated November 6, 2026. Related: Firecrawl review · Firecrawl alternatives · Firecrawl vs Crawl4AI · Scraping hub.
1. Who this pricing page is for
You already think Firecrawl might be the product. You need the bill. This page maps official hosted prices, credit rules, and five workloads onto a plan. It is not a second review: the purchase-fit question (“is a hosted LLM context API what I need?”) is answered in the Firecrawl review.
Use this page when…
- You searched “Firecrawl pricing” or “how much does Firecrawl cost.”
- You need to know whether Free, Hobby, or Standard covers your monthly pages.
- You want pay-as-you-go math before you over-upgrade.
Leave this page when…
- You are still deciding whether Firecrawl is the right category — start with the review.
- You want a landscape of other vendors — use Firecrawl alternatives.
- You want a pairwise duel — use the vs pages linked from that alternatives article.
UA did not create a billing account or generate an invoice. Every figure below is taken from Firecrawl’s official pricing page and billing documentation, checked November 6, 2026. Where those two pages disagree, both wordings are quoted. No success rate or speed claim is attached to a price.
2. Official plans and how credits work
Hosted figures from Firecrawl’s official pricing page, November 6, 2026 (USD).
| Plan | Monthly | Annual display | Credits / month | Concurrent browsers |
|---|---|---|---|---|
| Free | $0, no card | $0 | 1,000 | 2 |
| Hobby | $19 | $16/month ($190/year) | 5,000 | 5 |
| Standard | $99 | $83/month ($990/year) | 100,000 | 25 |
| Growth | $399 | $333/month ($3,990/year) | 500,000 | 50 |
| Scale | $749 | $599/month ($7,190/year) | 1,000,000 | 100 |
| Enterprise | Custom | Custom | Custom | Custom |
Credit spend, from the official pricing page and billing docs:
- Scrape or crawl: 1 credit per page.
- PDF parsing: billing docs list +1 credit per PDF page. The marketing pricing page also describes PDF cost as 1 credit per PDF page. We cite both and do not pick a winner.
- Map: 1 credit per call.
- Search: 2 credits per 10 results, rounded up. Eleven results = 4 credits. Scraping those hits adds the scrape cost on top.
- JSON,
question/query, andhighlights: +4 credits per page each. A JSON scrape is therefore 5 credits. Prompt-injection check, PII redaction, audio, and video also add 4 each and stack: JSON + PII = 9 credits. - Interact: pricing page lists 2 credits per browser minute. Billing docs: one-minute minimum;
promptsessions bill at 7 per minute; Playwrightcode-only at 2; zero-data-retention adds 2 on top. - Agent: 5 free daily runs, then dynamic pricing. Do not treat Agent as a flat per-page scrape.
- Enhanced proxy: same 1 credit as a basic request. An escalated retry is not charged separately.
- Social fetches called out in billing docs: typical X/Twitter via Grok around 30 credits (1 + 29); LinkedIn provider-routed fetches in a 5–40 credit range.
If Firecrawl returns a document, you pay — including a captured 403 or 404. If it returns no document, a basic scrape is not charged, with documented exceptions (a prompt-injection check that already ran, or interact time already consumed).
3. Real cost by scenario
Which published plan covers the work you actually run? Calculations use official rates checked November 6, 2026 for the workloads described — no retries, no extra modifiers unless stated. Hard, authenticated, or PDF-heavy pages move the number. These scenarios exist so you can pick a plan, not so you can compare Firecrawl to another vendor (that lives on the vs pages).
Scenario A — Prototype: 800 Markdown pages a month
Basic scrape is 1 credit per page, so 800 credits fit the Free plan’s 1,000. Cost: $0, with 2 concurrent browsers and a 10/min scrape cap. The same 800 pages with JSON extraction cost 5 credits each (billing docs): 4,000 credits, which needs Hobby at $19/month (or $16/month billed annually). If half the pages stay Markdown and half become JSON, you spend 800 × 0.5 × 1 + 800 × 0.5 × 5 = 2,400 credits — still Hobby, still not Standard.
Scenario B — Production Markdown: 10,000 pages a month
10,000 basic credits overflow Hobby (5,000) and sit inside Standard (100,000). Published step: Standard at $99/month, or $83/month billed annually. Concurrency jumps from 5 browsers to 25, and scrape throughput from 100/min to 500/min. Paying Hobby plus pay-as-you-go for the extra 5,000 credits would mean at least five $5 Hobby packs (1,000 credits each) on top of $19: $44, and you would still only have 10,000 credits with Hobby’s 5-browser cap. Standard is the cleaner buy at this volume even before you value throughput.
The same 10,000 pages as JSON: 50,000 credits. That is still Standard. JSON at 10k pages does not, by itself, force Growth.
Scenario C — Structured extraction: 50,000 JSON pages a month
50,000 × 5 credits = 250,000 credits. Standard’s 100,000 is not enough. Growth’s 500,000 is. Published step: Growth at $399/month, or $333/month billed annually. Could you stay on Standard and buy pay-as-you-go for the extra 150,000 credits? Standard packs are $5 for 2,000 credits, so 150,000 credits would take 75 packs: $375, plus the $99 plan = $474 — more than Growth’s $399, and you still would not get Growth’s 50 browsers or 5,000 scrapes a minute. Upgrade.
If those 50,000 pages stay Markdown (1 credit), they fit Standard’s 100,000 with room to spare. The plan jump is caused by JSON, not by “50,000 URLs” as a headline.
Scenario D — Search then scrape: 200 queries, 10 results, scrape the top 3
Search billing is 2 credits per 10 results, rounded up. Two hundred queries that each request 10 results cost 200 × 2 = 400 search credits. Scraping the top 3 hits is 200 × 3 = 600 basic scrape credits if you take Markdown, or 200 × 3 × 5 = 3,000 credits if you extract JSON. Combined Markdown path: 1,000 credits exactly — Free on paper, with no room for retries or 404 documents. Combined JSON path: 3,400 credits — Hobby. If each query requested 11 results instead of 10, search would round up to 4 credits per query (800 search credits) before any scrape.
Scenario E — Interact sessions: 40 minutes of browser time
Using the marketing pricing-page rate of 2 credits per browser minute, 40 minutes = 80 credits, which fits Free. Using the billing-docs prompt path at 7 credits per minute, 40 minutes = 280 credits, still Free. Add the documented one-minute minimum per session: twenty 90-second prompt sessions bill as 20 × 2 minutes × 7 = 280 credits, not 20 × 1.5 × 7. Add zero-data-retention at +2 per minute on the billing-docs wording and the prompt path becomes 9 credits per minute: 40 × 9 = 360 credits, or 20 sessions × 2 × 9 = 360. Interact only becomes a plan problem when it is a daily habit measured in hours, or when it sits on top of a large scrape job that already filled the bucket.
Agent is separate: 5 free daily runs, then dynamic pricing. There is no honest flat monthly number we can publish for Agent beyond that documented free slice. If Agent is the product you came for, treat the plan fee as access and expect a usage line you will only see on an invoice.
4. Pay-as-you-go versus upgrading
Pay-as-you-go on paid plans adds credits in $5 increments: 1,000 on Hobby, 2,000 on Standard, 2,500 on Growth, 5,000 on Scale. Bought credits stay until used and expire if you cancel. Free has no pay-as-you-go — billing docs say you get HTTP 402 when the bucket is empty.
Implied extra-credit prices if you only look at the pack size:
| Plan | $5 pack | Implied extra-credit price | When PAYG still makes sense |
|---|---|---|---|
| Hobby | 1,000 credits | $0.005 per credit | A one-off spike of a few thousand credits, not a new baseline. |
| Standard | 2,000 credits | $0.0025 per credit | Occasional overflow under ~40,000 extra credits; above that, run the Growth comparison. |
| Growth | 2,500 credits | $0.002 per credit | Spikes on top of 500,000, not a substitute for Scale. |
| Scale | 5,000 credits | $0.001 per credit | True overflow on a 1,000,000-credit baseline. |
A simple Hobby-to-Standard check: you are on Hobby ($19, 5,000 credits) and you now need 20,000 credits most months. Fifteen extra Hobby packs would be 15 × $5 = $75, plus $19 = $94 — almost Standard’s $99 — and you would still have Hobby’s 5 browsers and 100 scrapes a minute, not Standard’s 25 and 500. At 20,000 credits, Standard also leaves 80,000 credits of headroom for JSON or retries. Upgrade.
A Standard-to-Growth check already appeared in Scenario C: 150,000 extra Standard credits via PAYG cost $375 + $99 = $474 versus Growth at $399. The upgrade wins on money and on limits.
Annual billing is a separate lever. Hobby $190/year versus 12 × $19 = $228; Standard $990 versus 12 × $99 = $1,188; Growth $3,990 versus $4,788; Scale $7,190 versus $8,988. Annual is cheaper if the plan is stable. It is a bad prepay if you are still discovering that JSON will kick you up a tier next month.
5. Rate limits, concurrency, and rollover
Official rate limits, Free through Scale:
| Family | Free | Hobby | Standard | Growth | Scale |
|---|---|---|---|---|---|
/scrape, /map, /search per minute | 10 | 100 | 500 | 5,000 | 10,000 |
/crawl and /agent per minute | 2 | 20 | 100 | 1,000 | 2,000 |
/interact per minute | 2 | 20 | 100 | 1,000 | 1,500 |
| Concurrent browsers | 2 | 5 | 25 | 50 | 100 |
Throughput can force an upgrade even when credits would have been enough. A Hobby job that needs 30 pages in flight is a concurrency buy, not a credit buy. Crawl is the tightest everyday cap: 2/min on Free is a prototype pace; 20/min on Hobby is still modest for a large docs domain.
Unused plan credits do not roll over on Free, Hobby, Standard, or Growth. The pricing table lists Scale rollover as one month; billing docs specify that for annual Scale (one month) and annual Enterprise (two months). Bought pay-as-you-go credits persist until used and expire if you cancel. That split matters: a Standard subscriber who ends the month at 40,000 unused plan credits loses them; leftover PAYG packs do not reset on the same rule.
6. Billing pitfalls
- Budgeting “pages” instead of credits. JSON, search, interact, stacked +4 modifiers, and social fetches all break a 1:1 mental model.
- Paying for returned error documents. A captured 403 or 404 is billed. Retrying a wall without changing approach can empty Free or Hobby without producing Markdown.
- Using Free as if it had a top-up. It does not. HTTP 402 is the documented next step.
- Buying PAYG out of habit instead of upgrading. The two worked examples above (Hobby → Standard around 20,000 credits; Standard → Growth around 250,000 JSON credits) are the common places that habit gets expensive.
- Prepaying annual on a plan you will outgrow once JSON or Agent leaves evaluation.
- Ignoring the Interact discrepancy. If you use prompt sessions, budget 7 credits per minute (billing docs), not 2 (marketing page), and count the one-minute minimum.
- Treating Agent’s 5 free daily runs as a production allotment. After that, pricing is dynamic.
7. Which plan to pick
| Monthly shape | Published plan to start from |
|---|---|
| Under 1,000 basic Markdown credits, low concurrency, no need for top-ups | Free ($0) |
| Up to about 5,000 credits, including some JSON, or you need 5 browsers / 100 scrapes a minute | Hobby ($19, or $16/month billed annually) |
| Roughly 5,000–100,000 credits, production Markdown or moderate JSON, or 25 browsers | Standard ($99, or $83/month billed annually) |
| Above 100,000 credits — typical once large JSON jobs appear — or 50 browsers | Growth ($399, or $333/month billed annually) |
| Toward 1,000,000 credits, 100 browsers, or you want the documented Scale rollover | Scale ($749, or $599/month billed annually) |
| Custom contracts, extra retention, or volumes past Scale | Enterprise (custom; ask Firecrawl, not us) |
Pick the plan for the job you will run next month, measure credits on real pages, then either buy a small PAYG pack for a spike or step up a tier. If the numbers above already feel like the wrong category of expense, stop and reread the review: you may want a different tool, not a different Firecrawl SKU.
8. FAQ
How much does Firecrawl cost per month?
Published hosted prices on November 6, 2026: Free $0; Hobby $19 (or $16/month billed annually); Standard $99 (or $83); Growth $399 (or $333); Scale $749 (or $599). Enterprise is custom. Your invoice then depends on credit modifiers and any $5 pay-as-you-go packs.
Does Firecrawl have a free plan?
Yes. 1,000 credits a month, no credit card, 2 concurrent browsers. No rollover and no pay-as-you-go. Billing docs: HTTP 402 when credits run out.
Do unused Firecrawl credits roll over?
Not on Free, Hobby, Standard, or Growth. Scale lists one-month rollover on the pricing table; billing docs attach that to annual Scale (one month) and annual Enterprise (two months). Purchased PAYG credits persist until used and expire on cancel.
How many credits is a JSON scrape?
5, according to the billing docs: 1 for the scrape plus 4 for JSON. Other +4 modifiers stack on top.
Is annual billing worth it?
Yes if you already know which tier you will stay on: Hobby $190 vs $228, Standard $990 vs $1,188, Growth $3,990 vs $4,788, Scale $7,190 vs $8,988. No if JSON or volume will push you up a tier within a month or two.
Where do I go if I am not sure Firecrawl is the right tool?
The Firecrawl review for fit; Firecrawl alternatives if the job might belong to another category.
9. Sources
Vendor pages are named in text so this article does not send click-throughs to non-affiliate destinations. Dates are Europe/Paris verification dates.
- Firecrawl official pricing page (checked November 6, 2026)
- Firecrawl official billing documentation (checked November 6, 2026)
- Firecrawl official documentation (checked November 6, 2026)