Roundup
Free Web Scraping API, or Just a Free Trial? How to Tell
Some free scraping APIs renew every month; others are one-time trials. What each free tier really covers once JavaScript and billable errors count.
I would choose a free scraping API by whether the allowance repeats and covers the required feature, not by the largest signup number. ZenRows is a shortlist for recurring generic credits; Firecrawl fits a monthly scrape/crawl path; Bright Data advertises recurring Unlocker requests; ScrapingBee and Zyte belong in the trial category.
The limits below come from public documentation; no accounts were acquired or requests executed for this article. A credit, a dollar and a browser minute cannot be ranked as the same amount of useful output.
Short on time?
Here's my take
- Recurring generic credits
- ZenRows Free
- Inspect the documented 5K monthly credits and feature weights; the agent storefront is a separate one-time offer.
- Scrape/crawl output
- Firecrawl Free
- Its 1K monthly credits need endpoint/modifier and rate-limit checks, not a universal 1K-page promise.
- Small prototype trial
- ScrapingBee or Zyte
- Use the documented signup/trial window as an evaluation budget; recurring access was not established.
Prices verified . Public offers checked 11 November 2026. No grant, quota, billing or extraction quality was tested.
The free offers, side by side #
| Provider | Documented free phase | What the unit buys | Important limit/condition |
|---|---|---|---|
| ZenRows | 5K credits/month | Fetch/JS/proxy features use different total weights | Free concurrency 5; storefront 5K once is a different path |
| ScraperAPI | 5K credits for 7 days, then 1K/month | Endpoint/domain/feature credit totals | No-card trial; documented ongoing concurrency 5; free-feature rights unverified |
| Firecrawl | 1K credits/month; no card | Base scrape one/page; other outputs/endpoints add costs | Free concurrency 2 and endpoint-specific rates |
| Bright Data Unlocker | Advertised 5K requests/month | Its successful-request product | Account/target eligibility and exhaustion untested |
| Scrapingdog | 100/month | Generic/JS/proxy weights differ | One concurrency |
| ScrapingBee | 1K signup credits; no card | Plain/default-JS and feature totals differ | Recurring grant not established |
| Zyte | $5/30-day trial once | Target-tier HTTP/browser requests and other charges | Not a fixed recurring page count |
| Crawlbase | New-account 1K, up to 5K after setup | Target/JS billing path | Recurring grant not stated |
| Apify | $5/month spend budget | Actor compute, proxy/storage and other fees | Not five dollars worth of a universal page rate |
Checked 11 November 2026. I label trial/setup phases separately from monthly allowances; rows do not establish a matched rendered-page or usable-record winner.
Which free tiers renew every month? #
ZenRows Free publishes 5K monthly credits; its agent storefront instead describes a one-time grant after email verification. I would keep those purchase paths separate rather than turn both into an ongoing balance. Plan pricing, agent payments, plan FAQ.
ScraperAPI describes 1K monthly credits/five concurrency after its seven-day 5K trial. I would check the actual free feature rights before relying on a JS or geography workflow. Pricing, billing FAQ.
Firecrawl’s monthly offer is 1K credits without a card, with two concurrency. Its base scrape page is one credit, but other formats/features and billable returned error pages change the usable amount. Pricing, billing. Paid plans and credit rules are covered in Firecrawl pricing, explained.
Bright Data advertises 5K monthly Unlocker requests, while Scrapingdog’s table lists 100/month and one concurrency. These are documented options to inspect for the matching product, not a shared entitlement or measured target-success comparison. Unlocker, Scrapingdog. The web unblocker comparison covers what Unlocker costs once the free requests run out.
Which ones are just trials? #
ScrapingBee’s documented 1K-credit offer is a signup trial, not a recurring allowance established here. Its current pricing/help define the trial and feature weights; the general FAQ adds account context. I would plan a finite evaluation rather than a monthly production budget from that number. Pricing, trial help, FAQ.
Zyte’s $5/30-day trial is one-time, with HTTP/browser prices depending on target tier. Crawlbase describes a new-account allowance that grows after setup, not a stated recurring monthly grant. Zyte, target-tier billing, Crawlbase.
My recommendation is to evaluate the exact required mode during the documented trial, preserving actual outcomes separately from a sales promise. No trial was started for this article.
How far do free credits go with JavaScript? #
Apply the feature total before estimating pages. ScrapingBee plain mode is one credit with render_js=false; default JS is five. ZenRows Fetch plain/JS weights are one/five, and ScraperAPI generic JS is ten. Bee docs, ZenRows pricing, ScraperAPI JS.
These are hypothetical upper bounds before other features, special targets or billable errors:
| Documented balance | Plain at one credit | Generic JS at the selected weight |
|---|---|---|
| Bee 1K signup credits | Up to 1K requests | Up to 200 at five credits |
| ZenRows 5K monthly credits | Up to 5K requests | Up to 1K at five credits |
| ScraperAPI 1K monthly credits | Up to 1K requests | Up to 100 at ten credits, if free feature rights fit |
A request is not a validated useful record. ScraperAPI and Bee can bill specified non-2xx outcomes, and Firecrawl can bill returned error-page content; I would track actual evidence before calling the balance a useful-page guarantee. ScraperAPI costs, Bee status rules, Firecrawl billing.
What about dollar budgets and browser minutes? #
Apify’s recurring $5 is a spend allowance. Its compute unit is allocated 1GB RAM for one hour, and the chosen Actor/proxy/storage fees determine consumption. I would not convert that balance into a universal free-page number. Pricing, resources. The Apify pricing guide shows how compute units and Actor fees add up past the free budget.
Cloudflare Browser Run’s Free allowance is ten minutes daily, with three browsers, rather than a scraping-request credit pool. I would choose it for the documented browser job only when its daily distribution and limits fit. Pricing. What happens after the free minutes is in my Cloudflare Browser Run pricing breakdown.
The scraping credits explainer shows how to compare credits, money, time and output without treating them as identical units.
Free credits don’t give you permission #
A grant does not give target/data rights. ScrapingBee terms require customer authorization and an AUP, while ZenRows terms restrict unlawful uses and control free eligibility. I would review the applicable targets and contract before a future run. Bee terms, AUP, ZenRows terms.
Paid renewal, overage and credit exhaustion depend on the selected path. The offer table does not establish a private account hard stop or authorize adding funds; no payment setting was inspected or changed here.
I would decide the allowed test, output acceptance and spend boundary before using the balance. That is a proposed review checkpoint, not a free-key experiment performed for this post.
Which offer should you choose? #
Pick a recurring allowance only when its product/feature scope and limits fit the intended ongoing work. Pick a trial/setup grant for a finite prototype, and keep Apify’s dollar budget or Cloudflare’s minutes separate from generic credits.
I would shortlist ZenRows for its documented monthly generic balance, Firecrawl for its scrape/crawl route, or a smaller recurring offer when it already matches the application. For a paid decision, use ScraperAPI pricing, ScrapingBee pricing and ZenRows pricing.
The scraping API hub connects these free offers to the paid choices.
FAQ #
Is the biggest free signup offer the best deal? #
Not automatically. I would distinguish recurring allowances, one-time trials and setup conditions before applying feature weights.
Does ScrapingBee give 1,000 free credits every month? #
A recurring grant was not established in this research. The documented offer is a signup trial. Pricing, trial help.
Do free credits mean free usable data? #
No. Feature weights, eligibility and billable error/cancellation rules can reduce the records that survive validation.
Does a free plan let me scrape any site? #
No permission was established that way. Review the actual target rights, AUP and terms for the chosen account/path. Bee terms, ZenRows terms.
Sources
24 sources, accessed .
Official pricing, docs, changelogs and status pages 24
- ScraperAPI plans
- Request credit costs
- Plans and billing
- JS rendering FAQ
- ScrapingBee plans
- HTML API
- ScrapingBee FAQ
- Service terms
- Trial allowance
- Acceptable use
- Zenrows monthly pricing
- General FAQ
- Agent storefront
- Zenrows service terms
- Zyte pricing
- Zyte API billing
- Apify pricing
- Actor resources
- Firecrawl pricing
- Firecrawl billing
- Web Unlocker pricing
- Browser Run pricing
- Scrapingdog pricing
- Crawlbase pricing
Last updated: 13 October 2026. Researched and drafted with AI tools; human editorial review and publication approval are pending.
