Cheapest Web Scraping API: Costs by Volume and Mode

• Zyte API: 2026-10-31 (live pricing page; also the same-day fact-checks of cheapest-web-scraping-api and scraperapi-alternatives). Tier 1 (easiest sites) prices. Pay-as-you-go: HTTP $0.13–1.27 and browser $1.01–16.08 per 1K across tiers 1–5; the docs set a standard PAYG spend ceiling of $100. $100/mo commitment: HTTP from $0.10, browser from $0.75; $200: $0.08 / $0.60; $500: $0.06 / $0.48. JS at 100K uses the $100 commitment ($75 of usage, $100 floor); JS at 1M uses $500. Residential IPs have 'different base costs' that are not published, so premium is blank.
• ScrapingAnt: The browser parameter defaults to true, so a request with no options costs 10 credits ($1.90 per 1K on Enthusiast); plain figures assume browser=false. Plans: Enthusiast $19 (100K), Startup $49 (500K), Business $249 (3M), Business Pro $599 (8M), Custom from $699 (10M+). Restarting Startup once in a month would make 1M plain $98 ($0.098) and 100K JS $98; I used single plans. JS and premium at 1M need Custom, so they are blank.
• scrapestack: Basic $19.99 (200K; no JS, no premium proxies; +$0.0003998 per extra request), Professional $79.99 (1M; JS and premium proxies; +$0.00031996), Business $199.99 (3M; +$0.0002666533). How many requests a JS or premium-proxy call uses is not on the pricing page, and the docs sit behind an APILayer page I could not read, so JS and premium are blank.
• Scrape.do: Plans: Hobby $29 (250K), Pro $99 (1.25M), Business $249 (3.5M), Advanced $699 (10M). The docs tell you to trust the Scrape.do-Request-Cost response header, because some domains have server-side defaults.
• ScrapeOwl: Plans: Bootstrap $29 (250K), Startup $99 (1M), Business $249 (3M). JS and premium at 1M need more than 3M credits, so they are blank.
• ScrapFly: 2026-10-16 pricing + billing doc: Discovery 30/200K, Pro 100/1M (+3.50/10K), Startup 250/2.5M (+2/10K) and Enterprise 500/5.5M (+1.20/10K). Human-readable pricing says Discovery has a hard cap; the machine-readable table also lists 5/10K but says quota-cap, so no Discovery overage is used. Credits reset at renewal. Datacenter base costs 1 and residential base 25; rendering adds +5, giving 6/30. These base-credit examples assume HTML below bandwidth thresholds; large HTML >1MB, request body >100KB and rendered traffic >3MB incur bandwidth credits. Failed-request protection excludes named 4xx codes and lapses above 30% eligible failure traffic/minimum hour. JS at 1M=560, premium at 1M=2840 with Enterprise + overage. Public-price arithmetic only, no performance/API tests.
• ZenRows: 2026-10-31 (live docs pricing page; also the same-day fact-checks of cheapest-web-scraping-api and scraperapi-alternatives) monthly rungs: Build 19/45K, 29/80K, 39/120K; Launch 69/250K, 99/500K, 129/700K; Growth 199/1.2M, 279/2M, 399/3M; Scale 549/5M, 749/8M, 999/12.5M, 1999/28M, 2999/45M. Credit weights are 1/5/10/25. Free 5K monthly/no rollover. Plan credits expire at every renewal; unused top-ups roll into the next period once. Published top-ups by rung, with up to 10 purchases per period, were considered; existing displayed volume totals remain cheapest for these targets. Annual billing is charged upfront with two months free; on 16 Oct the annual view rounded entry prices to Build 16, Launch 58, Growth 166 and Scale 458 (the 458 Scale figure was seen again on 31 Oct); September 27 annual labels of 57/165/456 are historical. No API execution.
• Scrapingdog: 2026-10-31 (live pricing page; also the same-day fact-checks of cheapest-web-scraping-api and scraperapi-alternatives) plans: 40/200K, 90/1M, 200/3M, 350/6M, 500/9M and 1000/19M, up to 30000/1.1B monthly credits. The public table confirms generic 1, JS 5, premium 10 and combo 25. Premium at 1M needs 10M credits, so it uses Business Plus 1000/19M. Failed/blocked/timed-out requests are credited back; rollover starts from Standard. The FAQ says a downgrade resets credits to the new plan's allowance and plan changes are not prorated. The JS default is not stated in the inspected docs. Free-credit wording conflicts: the table puts 100 in a 'Monthly credits' row, the FAQ says every account 'starts with 100 free credits'; both readings are kept and the 100 is not counted as a recurring grant. The historical September 25 free grant of 200 became 100.
• ScraperAPI: 2026-10-31 (live pricing page and credit-cost docs; also the same-day fact-checks of cheapest-web-scraping-api and scraperapi-alternatives): Hobby 49/100K, Startup 149/1M, Business 299/3M, Scaling 475/5M, Professional 975/10.5M and Advanced 1975/21.5M credits. The last two carry limited-offer bonus labels (+250K, +500K); do not assume those bonuses recur forever. Generic plain costs 1, render 10, premium 10 and combo 25; ultra costs 30, or 75 with JS (paid plans only). Cost docs label modifiers 'Extra Cost', while the JS FAQ specifies 10 total; the normalized JS cells use 10 total. PAYG exists from Scaling at a fixed per-credit rate that is not published; no such overage is normalized. Credit balance resets at renewal. Free 1K monthly/5 concurrent and an initial 5K 7-day trial with no card are on the pricing page and Plans & Billing docs. No API execution.
• ScrapingBee: 2026-10-31 (live pricing page and API docs; also the same-day fact-checks of cheapest-web-scraping-api and scraperapi-alternatives): render_js defaults to true at 5 credits; plain figures explicitly disable it. Credits are 1/5/10/25; stealth_proxy costs 75. Monthly plans are 19/75K, 49/250K, 99/1M, 249/3M and 599/8M. Published Enterprise tiers are Enterprise 14 at 999/14M, Enterprise 24 at 1599/24M, Enterprise 41 at 2399/41M, Enterprise 69 at 3599/69M and Enterprise 120 at 5799/120M. Thus 1M premium requests (10M credits) use Enterprise 14 at 999 total, rounded to $1/1K. The pricing page says credits are valid for one month and don't roll over (previously recorded as not stated). Default billed codes are 200/404/410/413; transparent_status_code mode bills other below-500 codes too. Auto Mode costs 0 if every configuration fails. No API execution.
• Crawlbase: 2026-10-16 rate card: The first 1K costs $3, the next 10K $2, the next 100K $0.60, the next 1M $0.50, the next 10M $0.10, then $0.05/$0.04/$0.02 per 1K. At 100K, plain costs 76.40. The current page explicitly says difficulty and JS multiply a group's whole ladder total, so 1K JS=$6, resolving the old $5/$6 ambiguity. Subscriptions are 99/200K, 199/500K, 349/1M and 599/2M; usage beyond the block uses the same graduated ladder. At 1M plain costs 349; 100K JS costs 99; 1M JS costs 599. Free 1K initially, up to 5K after setup, not monthly. Only successful requests are billed; no proxy-mode equivalence is assumed.
• Bright Data Web Unlocker: 2026-10-31 (live pricing page and Web Unlocker feature docs; also the same-day fact-checks of cheapest-web-scraping-api and scraperapi-alternatives): Free Tier 5K requests/month; PAYG $1.50/1K successful requests; Scale $499/month for 383K requests plus $1.30/1K additional; Enterprise custom. At 1M, Scale costs $499 + 617 × $1.30 = $1,301.10, or $1.30/1K, cheaper than PAYG at $1,500. On 16 Oct the Scale slider also showed $999 for 908K then $1.10/1K and $1999 for 2M then $1/1K, which gave $1,100.20 ($1.10/1K) at 1M; those stops were not visible on 31 Oct, so the 1M cells now use the visible $499 card. Paid totals are gross of free allowances. Browser rendering is included; force rendering uses render=true. No selectable residential-only proxy mode was verified, so premium-mode comparison cells are withheld. The feature docs still describe opt-in Premium Domains, billed at a higher rate for listed domains only; the pricing page shows no premium-domain rate, and generic pages here use the default tier. Custom headers/cookies/expect features bill all attempts: https://docs.brightdata.com/products/web-unlocker/features . The page advertises a first-deposit match up to $500, which this dataset never counts. Free/paid stacking, actual performance and invoice behavior remain unverified.
• Scraping Fish: 2026-10-16 buy page: Packs are 2/1000/1 month, 20/10250/3 months, 200/105000/6 months and 2000/1.1M/12 months, plus VAT if applicable. Under this dataset's expiring-pack method, 100K monthly can consume the 1.1M pack within 11 months; 2000/1100=$1.818 per 1K is cheaper than 200/105=$1.905, so the 100K cells change to 1.82 and require 2000 upfront. At 1K monthly the cost remains $2/1K; at 1M monthly it is 1.82. JS and mobile proxy are included at the same flat request price; 25 concurrent/domain; 200/404 are billed. No execution; the September 27 browser read remains historical.
• HasData: 2026-10-31 (live pricing page; also the same-day cheapest-web-scraping-api fact-check): the page shows Startup $49/200K, Basic $99/1M and Growth $208/3M credits per month (concurrency 5/15/50) beside a 'Billed annually 2 months free' label, with Enterprise custom. On 16 Oct I recorded monthly-billed prices of $59/200K, $119/1M and $249/3M plus $415/5M, $830/10M and $1660/20M tiers, reading $49/$99/$208 as annual-purchase monthly equivalents. Neither the $59-$1660 figures nor the larger tiers were visible on 31 Oct, so the monthly-billed price is unresolved and the entry price and every paid cost cell are withheld; both readings are kept here. Shared credits cost 1 plain, 10 JS, 5 residential or 15 combined, charged only for success. The 1K plain cell stays 0 because the Free plan gives 1,000 credits every month. Self-serve request docs say unused credits expire; pricing-page rollover belongs to custom Enterprise. No account or invoice check.
• Oxylabs Web Scraper API: 2026-10-16: Public /Scrape pricing supersedes the old Micro 49/Starter 99/Advanced 249. Current plans are Starter 30, Basic 100, Flexible 500 and Enterprise from 2500. Generic /Scrape headlines start from 0.50/0.45/0.35/0.25 per 1K; Other-target prices start from 1/0.90/0.80/0.75. Google Non-JS at 1/0.90/0.70/0.50 and Google JS at 1.35/1.30/1.15/0.90 are target-specific examples, not a generic HTML tariff. All old exact mode totals/concurrency are withdrawn; target-specific current costs require a fixed configuration and supported tariff. No purchase or API usage.
• Decodo Web Scraping API: 2026-10-16: New Web Data API rates per 1K, for plan prices [19,49,99,249,499,999,1499], are standard [0.50,0.40,0.30,0.20,0.18,0.16,0.15], standard JS [0.75,0.65,0.60,0.55,0.50,0.45,0.40], premium [1,0.75,0.70,0.65,0.60,0.55,0.50] and premium JS [1.50,1.25,1.10,1,0.95,0.90,0.85]. proxy_pool defaults to premium; leaving headless unset disables JS. Current cards show 19 for up to 38K standard pages, 49 for up to 122K, 99 for up to 330K and 249 for up to 1.2M. Old alternative capacities of 38K/163K/707K are no longer current; normalized totals are withheld pending complete balance/top-up/free-grant matching. Public billing docs only; no account/API proof.
• Scrapeless Web Unlocker: 2026-10-16: The official Web Unlocker product page confirms Basic consumption-based pricing without a monthly commitment, base pricing from $1/1K requests and successful-only billing. Exact higher-tier commitments and discounts remain unverified, so normalized cheapest-volume cells stay withheld; old Growth/Scale/Business tariffs are not current claims. The pricing FAQ says unused monthly commitment is not retained: https://www.scrapeless.com/en/pricing . No account, vendor or product API action.

17 of 23 rows from Scraping API pricing tracker. Download CSV (CC BY 4.0). Scroll sideways for more columns.

• Crawl4AI Cloud: 2026-10-31 (live price API at api.crawl4ai.com/v1/prices; also the same-day fact-checks of crawl4ai-vs-firecrawl, firecrawl-pricing, firecrawl-alternatives, website-to-markdown-for-llms and cheapest-web-scraping-api): Cloud subscriptions replace the September 27 pack model: Starter 10/month for 10K credits +1/1K extra; Builder 99/125K +0.79/1K extra; Scale 499/750K +0.67/1K extra. Monthly usage caps are 50K/500K/2M credits and top-ups start at $10. Plain costs 0.5 credits, real browser 1, hard site 2, archive 0.25 and search 1. 20 guest credits need no signup; email verification then gives 1K credits a month, which covers 1K plain or browser pages. Plain at 100K: Starter 10 + 40K extras at 40 = 50 (exactly Starter's 50K cap); plain at 1M: Builder 99 + 375K extras at 296.25 = 395.25 (exactly Builder's 500K cap). JS at 100K: Builder 99/125K, because 100K credits exceed Starter's cap; JS at 1M: Scale 499 + 250K extras at 167.50 = 666.50. Rounded per 1K: 0/.50/.40 plain, 0/.99/.67 browser. These model consumed credits; cash charged in a month can differ because top-ups are bought in $10 steps. The vendor calls this a soft launch, so prices and free credits can change. Open-source library/self-hosting costs are separate. No API calls beyond reading the public price endpoint.
• Olostep: 2026-10-31 (live pricing page; also the same-day fact-checks of cheapest-web-scraping-api, firecrawl-alternatives and website-to-markdown-for-llms): Monthly plans are 9/5K, 99/200K and 399/1M successful requests. Credit packs of 20/10K, 200/250K and 1000/2M are valid for 6 months; the page offers them to people who don't like subscriptions but does not state outright that no plan is needed. Plans count successful requests while packs and the FAQ count credits; the page implies one credit per successful request without stating it. The expiry-aware method makes 1K monthly use the 20 pack with 6K usable=$3.333/1K, not 9 monthly; 100K monthly uses 200/250K=$0.80/1K, not 99/200K. At 1M monthly the cost remains 399/1M=$0.399. Upfront pack payments of 20 and 200 matter; no assumption of monthly pack renewal. Posts that price Olostep on monthly plans only show $9 at 1K and $0.99 at 100K. The page says trial requests are JS-rendered on residential IPs and paid plans include everything below them; Starter lists 150 concurrent; the trial is 500 successful requests, no card, with renewal not stated. No execution.
• Firecrawl: 2026-10-31 (live pricing page and billing docs; also the same-day fact-checks of firecrawl-pricing, firecrawl-alternatives, firecrawl-vs-apify, crawl4ai-vs-firecrawl, firecrawl-vs-tavily and cheapest-web-scraping-api): the HTML pricing page marks these prices 'Effective September 4, 2026'. Free 1,000 credits/month, no card. Monthly Hobby $19/5K, Standard $99/100K, Growth $399/500K, Scale $749/1M; annual cards $16/$83/$333/$599 a month, billed $190/$990/$3,990/$7,190 a year. Hobby has 5 concurrent browsers. Base markdown and enhanced proxy cost 1 credit/page. $5 top-ups buy 1K/2K/2.5K/5K credits by plan. Rollover conflict, both kept: the pricing page says Scale rolls over one month and Enterprise a custom amount; billing docs say annual Scale 1 month and annual Enterprise 2 months. The per-1K cells assume no rollover. PDF conflict: pricing.md says 1 credit/PDF page, billing docs say base plus 1/PDF page. PDF-inclusive totals are not normalized; resolve before PDF work. Plain/enhanced totals exclude PDF/JSON/guard options. Comparable JS columns remain withheld. No API execution.
• Exa Contents: 2026-10-31 (live pricing page and billing guide; also the same-day fact-checks of cheapest-web-scraping-api, firecrawl-vs-tavily, firecrawl-alternatives and website-to-markdown-for-llms): Standalone Contents costs $1/1K pages per content type; text-only is one content type. The first-team monthly free balance resets to 10 without accumulation; the onboarding bonus does not expire. Under this dataset's rule, the free allowance counts only when it covers the whole volume, so the displayed 0/1/1 at 1K/100K/1M remains. Content retrieval does not expose a comparable proxy/render configuration, so JS/premium are N/A. No observed API results.
• WebCrawlerAPI: 2026-10-16: PAYG has no monthly fee, from $0.002/page; Professional is $99/month, from $0.0015/page; Business is currently $499/month, from $0.001/page. LLM-cleaned markdown rates are $0.006/$0.0055/$0.005 per page respectively. The former $29 Starter is not asserted as current. Only from-rates are established; included subscription allowance and exact target cost remain unverified, so normalized per-1K cells stay withheld. Successful requests are billed. Trial signup needs no card; its size is unstated. No account/API proof.

5 of 23 rows from Scraping API pricing tracker. Download CSV (CC BY 4.0). Scroll sideways for more columns.

About this dataset. Download CSV (CC BY 4.0). Scroll sideways for more columns.