Cheaper Firecrawl alternative: Cost per page

The cheaper Firecrawl alternative is the tool with the lowest cost per successful page on your workload, and self-hosted or bandwidth-billed tools can win on plain static pages. Failed Web Data API work isn’t billed.

Your Firecrawl bill grew faster than your page count, or scrape jobs keep failing on protected sites. A lower per-page price doesn’t guarantee a lower bill once retries, output surcharges and blocked pages are counted.

We run Wayfern’s browsers and APIs, so we see where failed work goes. After reading, you can price a job per successful page and match each workload to a tool.

Key Takeaways

• Convert credits, GB, browser hours and tokens to dollars per successful page before you compare any two vendors.
• Billed retries push cost per successful page above the list rate, so ask every vendor how it bills failed work.
• On plain static pages, a self-hosted crawler usually costs less than Wayfern’s list rate.

What does Firecrawl cost today, and what do you pay for?

As Scrape.do’s roundup of Firecrawl alternatives describes it, Firecrawl sells web data in credits, so what you pay for is the credits each endpoint call uses, drawn from a monthly plan allowance.

Plans change. Copy those figures on the day you build your model. A cheaper Firecrawl alternative is only cheaper against current numbers.

Firecrawl’s docs name the endpoints you spend credits on and hold the full current list. Third-party write-ups such as a DEV Community comparison mention /scrape, /crawl and /extract. They also mention an /agent endpoint.

Two steps turn a Firecrawl plan into a reference rate for any cheaper Firecrawl alternative:

1. Divide the plan price by its included credits to get dollars per credit.
2. Multiply that by the credits one call to an endpoint uses, as the docs state.

Assume a plan costs $100 for 20,000 credits. One credit then costs $0.005, and a 1-credit call costs $0.005.

Any cheaper Firecrawl alternative has to beat that baseline per successful page. Our Wayfern vs Firecrawl comparison puts both tools on that same basis.

How do you price a cheaper Firecrawl alternative per successful page?

Cost per successful page is total spend divided by the pages that returned usable content, and four levers move it: retries, output add-ons, the proxy tier and how often protected sites block you. A list rate only gives you cost per attempted page.

Lever	Wayfern Web Data API	Ask any vendor
Failed work	Not billed	Are failed or retried requests billed?
Output add-on	+4 per page	Does JSON or other special output cost extra credits?
LLM summary	+2 per page	Is summarizing a separate charge?
Proxy tier	Datacenter first, residential on block	Does a stronger proxy multiply the rate?

Retries are the first lever. A DEV Community write-up on a 10,000-page Firecrawl job puts it this way: “You retry. More credits gone.” Read Firecrawl’s billing docs and Spider’s pricing page for how each treats failed or retried requests.

Wayfern’s pricing page says failed Web Data work is not billed, and every response returns the exact credits charged.

Add-ons raise the per-page rate. Firecrawl’s billing docs list per-page credits for scrape and crawl, plus any extra credits for JSON or other add-on modes. Wayfern’s scrape costs 1 credit per page, the output add-on adds 4 per page, and the LLM summary adds 2.

For example, one POST to https://api.wayfern.com/api/v1/scrape with the output add-on is 1 + 4 credits at the published rates. The response reports the credits charged, so you can check the bill call by call.

The proxy tier sets the price of a retry. The Wayfern API reference says that with proxy: "auto", Wayfern starts on a datacenter exit and retries on residential when blocked.

Blocked pages shrink the number of usable pages without touching the rate card. Measure them by running each tool, including any cheaper Firecrawl alternative, against your own target sites.

The arithmetic is short. Assume N attempts, U usable pages and a per-attempt price R.

• If failed work is billed: cost per successful page = (N × R) ÷ U. At N = 10,000 and U = 8,000, that’s 1.25 × R.
• If failed work is not billed: cost per successful page = (U × R) ÷ U = R.

Add any per-page surcharges to R, then rerun both lines with the U you measured.

How do credits, GB, browser hours and tokens compare for a cheaper Firecrawl alternative?

Credits, GB, browser hours and tokens compare on one basis once you convert each to dollars per 1,000 pages or dollars per open-page hour.

	Self-hosted (Crawl4AI, Crawlee)	Wayfern
Billing unit	Your own servers	Credits at $0.0125
1,000 plain pages	No per-page fee	$12.50 (1,000 credits)
One open browser hour	Your hardware	$0.10 (8 credits)
One residential GB	Your proxy vendor’s rate	$4.00 (320 credits)
Failed work	Costs you compute	Not billed

Wayfern’s pricing page sets the list rate at $0.0125 per credit and charges 1 credit per scrape. A thousand scrapes is 1,000 credits, or $12.50.

Browser time costs 8 credits per open-page hour, metered per second with a 60-second minimum per page.

Residential bandwidth costs 320 credits per GB, so 320 x $0.0125 = $4.00 per GB. Datacenter egress is free.

On plain static pages, Wayfern does not beat a self-hosted tool on price. One roundup of Firecrawl alternatives describes Crawl4AI as a free, open-source, self-hosted option that you run and maintain yourself, so the per-page fee is zero and the cost moves to servers, proxies and your time.

Bandwidth-billed or low-cost hosted options also come in under a per-page list rate on simple pages. The same roundup points to Spider.cloud for fast, low-cost extraction from known URLs.

Wayfern’s unit prices pay off when pages block bots or need a real browser, because you pay only for Web Data work that succeeds and each response shows its credit charge.

Each tool fits a different page type:

• Crawl4AI or Crawlee: plain public pages when your team will run the servers and proxies.
• Spider: cheap extraction of clean content from known URLs.
• Firecrawl: public, static pages turned into markdown for LLMs, per a DEV Community comparison.
• Wayfern: protected pages and browser sessions, with search, captcha and MCP on one credit balance.

Is self-hosting the cheaper Firecrawl alternative?

A self-hosted Firecrawl alternative is cheaper per page when your volume is high and your targets rarely block, because you pay for servers instead of per-page fees.

	Firecrawl, self-hosted	Crawl4AI or Crawlee	Wayfern, hosted
License	AGPL-3.0	Apache-2.0	Hosted service
Per-page fee	None	None	Credits
Who fixes breakage	You	You	We do
Downloadable binary	Yes	Yes	No

Firecrawl’s open-source core is licensed under AGPL-3.0, so have someone read the license before you modify it or host it for other people.

The Crawl4AI and Crawlee repositories both list the Apache-2.0 license on their GitHub pages. Each LICENSE file is the place to check for any added conditions.

Crawl4AI is described as a free, open-source, self-hosted option for teams “willing to run and maintain it yourself.”

What you maintain is the cost that no license file lists:

• Docker images and upgrades
• Proxy supply and rotation
• Browser and driver versions
• Monitoring and retries
• Captcha handling

Self-hosting as a cheaper Firecrawl alternative means no per-page fee, plus your engineering hours and your proxy bills. On documentation sites, blogs and open product pages, those costs stay small and self-hosting wins.

When sites block datacenter IPs, proxies and captcha solving become the largest line item.

Wayfern is hosted anti-detect Chromium, and we don’t offer a downloadable binary for self-hosting. Our credit rates are on the pricing page.

• Firecrawl self-hosted fits if you want its API shape and can accept AGPL-3.0 terms.
• Crawl4AI or Crawlee fits high-volume public pages where you want full control and no per-page fee.
• Wayfern fits when pages block bots, or when you want search, captcha and MCP on one balance. Our Firecrawl comparison covers the differences.

Which cheaper Firecrawl alternative fits your workload?

The cheaper Firecrawl alternative depends on which of four workloads you run, and each one points to a different tool.

1. Plain public pages: use Crawl4AI or Crawlee self-hosted, or Spider or Jina Reader hosted. A roundup of Firecrawl alternatives says to choose Spider.cloud when you need “fast, low-cost extraction of clean content from known URLs.” Firecrawl itself fits where its markdown output is what you need. Wayfern isn’t the cheapest option here.

2. Protected pages: use the Wayfern Web Data API with proxy: "auto", or a vendor whose published docs cover protected pages. Run 100 of your own URLs through each candidate and divide spend by the pages that came back usable.

3. Logged-in or multi-step flows: use a held browser session, which is the Wayfern Browser API driven by Playwright over CDP. As Wayfern’s browser sessions page puts it, a request/response endpoint returns one page and keeps nothing, so a login followed by three clicks needs a browser that stays open.

4. Search plus fetch: use the Wayfern Search API (Google, Bing and DuckDuckGo, 2 credits per provider results page), the Web Data API and the MCP server on one balance, with one bill and one integration instead of three. If you already run Firecrawl and your query volume is light, staying on its stack avoids a second vendor.

Our buyer’s guide to web scraping tools compares these options by billing unit.

How do you scrape logged-in pages with a hosted browser?

You scrape logged-in pages with a hosted browser by connecting Playwright to Wayfern’s anti-detect Chromium over CDP and signing in to an account you own or are authorized to use, within the site’s terms. Playwright’s documentation says connectOverCDP() “attaches Playwright to an existing browser instance using the Chrome DevTools Protocol.”

import { chromium } from "playwright";
import fs from "node:fs";

const browser = await chromium.connectOverCDP(
  "wss://api.wayfern.com/ws?token=LINK_TOKEN&country=US&region=CA&session=run42",
  { timeout: 120000 },
);
const context = browser.contexts()[0];
if (fs.existsSync("cookies.json")) {
  await context.addCookies(JSON.parse(fs.readFileSync("cookies.json", "utf8")));
}
const page = await context.newPage();
await page.goto("https://example.com/login");
// sign in, then collect the pages you are authorized to read
fs.writeFileSync("cookies.json", JSON.stringify(await context.cookies()));
await browser.close();

1. Connect with a link token; opening the socket starts the session.
2. Choose the exit. Set country, optionally region, and a session id such as run42. The sticky session holds one residential IP, so a cookie issued to one IP isn’t replayed from another. An unknown location closes the socket with code 4007.
3. Restore, sign in and save: your code loads cookies and storage at the start and writes them back before closing.
4. Close the browser when the run ends, because open pages are billed per second.

If a run stays logged in, page.goto() on an account page returns the account view instead of the login form.

If a person hand-manages many identities, Multilogin fits that job better than a hosted browser.

Run this snippet with a link token on an account you control.

How do you keep a scraping budget from running away?

A scraping budget stops running away when the balance, not an invoice, decides when work ends. Wayfern’s pricing docs put it this way: “Hard stop at zero. Nothing runs on credit you did not buy. Auto top-up is off until you turn it on.”

At zero, calls are refused and open sessions are closed.

Credit packs set a fixed spend:

• 1,000 credits cost $12.50, 5,000 cost $55 and 20,000 cost $200.

• Packs never expire and roll over.

• Packs work without a plan.

  Traffic through your own proxy isn’t metered as residential bandwidth, according to Wayfern’s pricing page. Captcha costs 1 credit per successful solve.

One balance covers the Browser API, Web Data API, Search API, captcha and MCP, so you keep fewer vendor bills and fewer integrations.

A mixed stack of specialists can still cost less on a single heavy workload, such as a large volume of plain static pages.

Frequently asked questions

What is the cheapest Firecrawl alternative?

The cheapest option depends on the page. A roundup of Firecrawl alternatives notes that for plain static pages, a self-hosted crawler such as Crawl4AI or a low-cost hosted API such as Spider can cost less per page. For blocked sites, logged-in flows, or search plus scrape on one bill, compare cost per successful page, which is where Wayfern fits.

Is there a free or open-source Firecrawl alternative?

Crawl4AI is a free, open-source, self-hosted option, and you run and maintain it yourself. Wayfern has no free tier or free trial.

Do failed requests cost credits?

On Wayfern’s Web Data API, failed work isn’t billed, and every response returns the exact credits charged, according to Wayfern’s pricing docs. Firecrawl and Spider set their own retry and failure billing rules in their billing docs, so read those before you compare totals.

Does Wayfern work with MCP?

Yes. You connect with a Bearer wf_ token that has the mcp scope, and calls bill at the same rates as the REST API. Two of the tools, list_search_engines and crawl_status, are free.

What should you do next to cut your scraping bill?

Measure your current cost per successful page first, then move each workload to whichever tool fits it, Firecrawl or a cheaper Firecrawl alternative.

Start today by pulling a week of Firecrawl usage and dividing total dollars by the pages that returned usable content. Any cheaper Firecrawl alternative has to beat that figure.

This week, run your plain static pages through Crawl4AI or Spider. Keep blocked, logged-in and search jobs in a separate queue.

This month, if blocked pages or sessions drive your bill, test them on Wayfern. Buy the smallest of Wayfern’s credit packs, 1,000 credits for $12.50, which never expires and works without a plan.

Run that first balance against your hardest pages and price them from the credits each response reports. Spending stops at zero, so the pack caps your risk.

Sources

Pricing | Wayfern Docs · Top 7 Firecrawl Alternatives for Web Scraping in 2026 · 6 Firecrawl Alternatives for Developers in 2026 - DEV… · Pricing - Credits for Browsers & the Web Data API | Wayfern · Wayfern API reference · 8 Best Firecrawl Alternatives in 2026: Real Reviews &… · Anti-Detect Browser Sessions over CDP | Wayfern · Wayfern MCP docs. Every other source is linked where the article cites it.