ZenRows pricing

About

ZenRows is a web-data infrastructure company whose platform bundles four “primitives” under a single subscription: Fetch (a request-based scraping endpoint with anti-bot bypass), Extract (auto-detected JSON parsing built on Fetch, currently in beta), Batch (large async web-data jobs with queues, retries, and webhooks, currently in beta), and Browser Sessions (a managed headless-browser session for pages that need clicks, forms, or login state). A legacy standalone Residential Proxies product remains available. All primitives draw from one shared monthly credit balance, accessible from every interface — API, SDK, CLI, and MCP server — so a customer never has to size separate quotas per product. The company markets anti-bot bypass for Cloudflare, DataDome, Akamai, PerimeterX, and reCAPTCHA as its core differentiator.

ZenRows targets developers, data teams, AI-agent builders, and scraping-heavy businesses across e-commerce, real estate, SERP, lead generation, price monitoring, and LLM-training data collection. Its pricing page claims more than 2,000 companies rely on its infrastructure and positions the product as built for “the web that fights back” — protected sites, browser workflows, recurring jobs, and high-volume pipelines — rather than competing purely on raw request volume.

The pricing structure is a credit-based tiered monthly subscription with five named tiers — Free, Build, Launch, Growth, and Scale — each split into three price/credit “rungs” (e.g., Build B1/B2/B3), capped by a custom Enterprise tier for usage above 12.5M credits/month. This replaces the company’s earlier “Regular”/“Business” tab structure and its Universal Scraper API / Scraping Browser / Residential Proxies product naming.

Pricing summary : credit-based subscription billed only for successful results

ZenRows uses a tiered monthly subscription on a shared credit balance that is billed for successful results only — failed and retried requests, and even HTTP 404/410 responses (which count as successful completions), don’t consume credits. Five named tiers each ship three rungs of increasing credit allowance: Free ($0), Build (from $19/mo), Launch (from $69/mo), Growth (from $199/mo), and Scale (from $549/mo), topped by a custom Enterprise plan above 12.5M credits/month.

The model has these dimensions:

• Credit multipliers — every successful request draws credits by difficulty: 1 credit for a standard request, 5 for JavaScript rendering, 10 for Premium Proxies, and 25 for both together. These weights are fixed across every plan.
• Extract and Batch draw credits at the same underlying Fetch rate — Extract adds no cost during its beta beyond the Fetch call it uses, and Batch bills per successful task at the same weights.
• Residential Proxies (Legacy) and Browser Sessions — bandwidth meters at a flat 25,000 credits/GB on every plan; Browser Sessions add a flat 5 credits per minute of session time.
• Concurrency scales by tier (5 on Free up to 200 on Scale, 400–1,000+ on Enterprise), independent of which rung is purchased within a tier.
• Monthly vs. Annual billing — an Annual toggle discounts every rung’s monthly-equivalent price by 17% (e.g., Launch $69 → $57/mo); the marketing pricing page defaults to this Annual view.
• Top-Up Packs — fixed-price credit packs sized per rung (e.g., Launch L1: 50,000 credits for $19.50), capped at 4 per billing period.

What makes this different: instead of a dollar-denominated CPM and per-GB rate card, ZenRows now runs a single internal currency — credits — that every primitive (Fetch, Extract, Batch, Browser Sessions, legacy Residential Proxies) draws from at fixed, published weights, so the only thing that changes between plans is how many credits a customer is given, not what anything costs to run.

Pricing by product

ZenRows Plans (Free / Build / Launch / Growth / Scale) — Monthly billing

The docs pricing page also lists a “GB max” per rung — the hypothetical bandwidth ceiling if a plan’s entire credit allowance were spent solely on Residential Proxies/Browser Sessions bandwidth (credits ÷ 25,000), not a separate quota. It runs from 0.2 GB on Free up to 500 GB on Scale S3.

Annual billing discounts every rung’s monthly-equivalent price by 17% (confirmed on rung 1 of each tier): Build $19→$16/mo, Launch $69→$57/mo, Growth $199→$165/mo, Scale $549→$456/mo. The marketing pricing page (zenrows.com/pricing) defaults to this Annual view; the table above uses the docs.zenrows.com Monthly-billing prices.

Credit multipliers (Fetch, Extract, Batch)

These credit weights are fixed on every plan — a protected request always costs 25 credits regardless of tier. Extract draws exactly what the underlying Fetch call would, with no added cost during its beta; Batch bills per successful task at the same weights as Fetch.

Residential Proxies (Legacy) and Browser Sessions

Rule of thumb published in the docs: 1 GB = 25,000 credits = 1,000 protected pages.

Top-Up Packs

Up to 4 top-ups per billing period; unused top-up credits roll into the next period. Free doesn’t include top-ups — Free customers wait for the next monthly refresh or upgrade to a paid plan.

Sales motions across products: PLG / self-serve for Free through Scale S3 (all priced publicly, instant signup); sales-led for Enterprise (custom credit pricing above 12.5M credits/month).

Hidden costs : multipliers, GB burn, and the protected-page premium

ZenRows’ “pay only for success” promise removes the most common scraping surprise — paying for blocked requests — but the multiplier system can still inflate a bill far past the headline CPM. The two examples below show how the same plan price produces very different real costs depending on the page mix.

Archetype A — a price-monitoring team on the Business ($299) plan scraping mostly protected sites. They run 1.5M requests/month, and 80% of targets sit behind Cloudflare/DataDome, requiring JavaScript rendering + premium proxies (×25):

The lesson: on protected-heavy workloads, the ×25 multiplier — not the sticker price — determines the tier you actually need. A team that budgets off the $299 headline can land on a $2,999 plan once the page mix is real.

Archetype B — a lead-gen team on the Startup ($129) plan scraping mostly basic pages plus light residential-proxy use. They run 800K basic requests and pull 20 GB through residential proxies:

The lesson: because proxy GB and Scraper API requests draw from the same balance, heavy proxy use silently eats the request budget — the shared balance is convenient but makes it easy to under-size a plan.

Want to estimate your own ZenRows bill? Use the ZenRows pricing calculator to model your monthly cost based on your basic/protected page mix, residential-proxy GB, and Scraping Browser session hours. For the broader pattern, see our guide to usage-based pricing metrics and our explainer on credit-based billing.

Pricing evolution : from request quotas to credits to a shared dollar balance

ZenRows has rebuilt its pricing model three times since launch. It began in 2021 with flat request quotas (Starter/Growth/Scale), shifted to a credit-based model where features consumed credits at multiplied rates by late 2021, moved to a shared dollar balance with explicit CPM and per-GB metering across three products by mid-2026, and then, within two months, replaced that dollar balance entirely with a shared “credits” currency spanning four renamed primitives by August 2026.

Cadence

Tracked range: 2021-06–2026-08 (archived pricing snapshots). Quarters not listed had no archived pricing change detected; mid-range snapshots between 2022 and 2025 were not all preserved, so some intermediate moves may be unobserved.

Notable changes

• 2021-06 — Earliest Wayback snapshot: Starter $49 / Growth $499 / Scale $999 on flat request quotas, with a $11/GB residential-bandwidth slider (web.archive.org 20210622).
• 2021-12 — Plans repackaged to Developer $49 / Startup $99 / Business $249 on API-credit allowances; the standalone per-GB residential line was replaced by “Unlimited Bandwidth” (web.archive.org 20211205).
• 2023-03 — Comparison-grid redesign; Enterprise quoted at $999+/mo; banner announcing a ~1.1M euro funding round (web.archive.org 20230315).
• 2026-06 — Shared dollar-balance model live: Developer $69 / Startup $129 / Business $299 plus a high-volume Business tab (Business 500 $499 → Business 3K $2,999); CPM and per-GB metering reintroduced as explicit dimensions.
• 2026-08-04 — Shared “credits” currency replaces the dollar CPM/per-GB system across four renamed primitives (Fetch, Extract, Batch, Browser Sessions); Free becomes a permanent $0/mo plan (5,000 credits) instead of a 14-day $1-usage trial; Developer/Startup/Business/Business-tab is replaced by Build/Launch/Growth/Scale (3 rungs each, $19–$999/mo); Enterprise now starts above 12.5M credits/month (zenrows.com/pricing, docs.zenrows.com).

The model shifts in detail

ZenRows has now re-architected its billing three times. The 2021 move from flat request quotas to API credits let it charge differentially for expensive work — a JavaScript-rendered, premium-proxied request burns 25 credits versus 1 for a basic fetch — without publishing 25 separate SKUs. The mid-2026 move to a shared dollar balance briefly went further, collapsing three products (API, browser, proxies) into one wallet priced in CPM and per-GB dollars, while reintroducing the per-GB proxy pricing that the 2021 “Unlimited Bandwidth” era had hidden. Then, within two months, the 2026-08-04 rebuild reversed the unit but kept the pooling: prices are once again denominated in credits rather than dollars, but this time the balance spans four primitives (Fetch, Extract, Batch, Browser Sessions) instead of one API product, and Residential Proxies — briefly a first-class metered line item — is demoted back to a legacy standalone add-on. The net effect across five years: the metering unit keeps oscillating between dollars and credits, but the trend line is consistent — broader pooling, finer-grained multipliers, and a lower headline entry price ($49 in 2021 to $19 for Build B1 in 2026, though the credit allowance behind that price isn’t directly comparable across eras).

What’s unique : one wallet, success-only metering, and the 404-counts-as-success rule

One shared balance across four primitives. Rather than sizing separate quotas for Fetch, Extract, Batch, and Browser Sessions, ZenRows pools a single credit balance that drains to whatever a customer uses; a legacy standalone Residential Proxies add-on draws from the same balance at a flat rate. This is unusual among scraping vendors that typically sell proxies and APIs as separate line items, and it lowers the cognitive overhead of buying — at the cost of making it easy to under-size a plan when one dimension spikes.

Free is now a permanent $0 plan, not a time-boxed trial. The August 2026 rebuild replaced a 14-day, $1-usage trial with a permanent Free tier (5,000 credits/month, no card required, “everything unlocked”). Removing the countdown clock changes the evaluation dynamic: prospects can test integration depth on their own timeline instead of racing a trial expiration, which typically lowers abandoned-evaluation rates at the cost of some free-riding risk that a low, fixed monthly credit ceiling is meant to contain.

Pay only for success, with 404/410 explicitly counted as success. Failed and retried requests never consume balance. The deliberate choice to count HTTP 404 and 410 as successful (the request completed and returned a valid signal) is a sharp edge: it protects ZenRows’ margin on dead URLs while still being defensible — the customer did get an authoritative answer.

Difficulty-priced metering via multipliers. The ×5 / ×10 / ×25 multiplier ladder prices the hardness of a scrape, not just its count. A basic public page and a Cloudflare-protected page draw from the same balance but at wildly different rates, aligning price with the underlying infrastructure cost (residential IPs, headless rendering). This is a cleaner expression of the value-metric selection problem than a flat per-request fee.

Self-serve top-ups instead of hard overage SKUs. When a customer hits a limit, they buy a fixed-price Top-Up Pack sized to their rung (e.g., Launch L1: 50,000 credits for $19.50) rather than being auto-charged surprise overages or hard-stopped, up to 4 per billing period. The August 2026 rebuild replaced the earlier “15% more usage for 15% of the plan price” formula with these flat per-rung pack prices — still a middle path between credit-based billing and pure pay-as-you-go, just priced discretely instead of as a percentage rule.

Strengths & weaknesses

Billing UX : shared credit balance, success-only metering, and rung-based top-ups

ZenRows exposes a handful of named billing controls on its pricing page, docs, and dashboard:

• Monthly credits slider — a slider on the pricing page (ticked at Free / 120K / 700K / 3M / 12.5M / Custom) that jumps the highlighted “Recommended” card to the matching rung.
• Monthly / Annual toggle — switches the displayed price between full Monthly billing and the Annual view, which discounts every rung’s monthly-equivalent price 17% (e.g., Launch $69 → $57/mo); the pricing page defaults to Annual.
• Shared credit balance — one credit balance spendable across Fetch, Extract, Batch, Browser Sessions, and legacy Residential Proxies, from any interface (API, SDK, CLI, MCP).
• Pay-only-for-success metering — failed and retried requests don’t consume credits; HTTP 404 and 410 responses count as successful.
• Top-Up Packs — fixed-price credit packs sized per rung (e.g., Launch L1: 50,000 credits for $19.50), capped at 4 per billing period, with an option to auto-trigger at a chosen usage threshold (e.g., 95%).
• Usage alerts — notifications at 80% and 95% of the monthly limit, configurable in Notification Settings.
• Rollover — unused monthly plan credits carry into the next period once, on every paid tier (Free has none).
• Upgrade / downgrade rules — upgrades apply immediately and carry over unused credits (base allowance, top-ups, and admin-granted credits) as bonus credits on top of the new plan’s allowance; downgrades apply at the next billing cycle, keeping current limits until renewal.

Strategic wins : pricing decisions that strengthened the model

1. Success-only billing turns reliability into a pricing promise

By charging only for successful requests, ZenRows converts its core technical claim — anti-bot bypass that works — into a billing guarantee a buyer can feel. A blocked request costs nothing, so the customer’s interests and ZenRows’ margin both point at higher success rates. This is a sharper alignment than per-request pricing, and it echoes the outcome-aligned thinking in our usage-based pricing fundamentals guide.

2. The shared balance lowers the buying decision to a single number

Pooling one credit balance across four primitives removes the hardest part of buying infrastructure: forecasting per-primitive usage. A buyer picks a monthly number and lets spend flow where it’s needed. That simplicity is a genuine conversion lever for self-serve PLG, the same dynamic we cover in our analysis of credit-based billing.

3. Difficulty multipliers price hardness without SKU sprawl

Instead of publishing dozens of SKUs for proxy types and rendering modes, ZenRows expresses cost as a ×1/×5/×10/×25 ladder over one base credit rate. It keeps the pricing page legible while still charging more for expensive work — a clean answer to the value-metric selection problem.

4. Public pricing to $999 removes friction across the self-serve range

Most infrastructure vendors gate high-volume tiers behind sales. ZenRows publishes its entire Free-through-Scale ladder — up to Scale S3 at $999/mo and 12.5M credits — as instant self-serve, capturing mid-market buyers who would otherwise bounce at a “Contact sales” wall. The August 2026 rebuild lowered this self-serve ceiling from the old $2,999/mo Business 3K plan to $999/mo, while moving where Enterprise custom pricing kicks in from a dollar threshold to a credit threshold (above 12.5M credits/month) — a gate that now scales with usage rather than sitting at an arbitrary price point.

5. Making Free permanent removes signup friction

Converting Free from a 14-day, $1-usage trial into a permanent $0/mo plan (5,000 credits/month, no card required) removes the countdown-clock anxiety that pushes trial users to over-provision or abandon evaluation early. Pairing a permanent free tier with usage-based upgrade triggers is a well-established PLG pattern, discussed further in our usage-based pricing fundamentals guide, and it costs ZenRows little marginal risk since Free’s credit ceiling is low enough to cap actual scrape volume.

Areas to improve : where the model creates avoidable confusion

1. Surface the effective price-per-credit by tier

Publishing a flat 25,000-credits/GB rate for Residential Proxies (Legacy) and Browser Sessions closed the old “derive your own $/GB” gap, but the credits-per-dollar exchange rate is still invisible: Build B1 works out to roughly $0.42 per 1,000 credits while Scale S1 is roughly $0.11 per 1,000 credits — a built-in volume discount of nearly 4x that a buyer can currently only find by dividing price by allowance themselves. Surfacing an effective ”$ per 1,000 credits” column next to each rung would make the volume economics as legible as the credit-per-request multipliers already are.

2. Add a worked “protected-heavy” example to the pricing page

The #1 third-party complaint is confusion over how charges break down. A single worked example on the pricing page — “1M protected requests on Growth G1 = X% of your credit balance via the ×25 multiplier” — would pre-empt the surprise of landing on a far larger plan than the headline suggested. The docs have the formula; the pricing page needs the punchline.

3. Clarify the 404/410-as-success rule before signup, not after

Counting dead URLs as successful results is defensible but counter-intuitive, and it lands worst for customers scraping volatile or expired URL sets. Stating it prominently at signup (rather than only in the FAQ) would protect trust. Pairing it with a dashboard breakdown of “billed-as-success but 404/410” responses would let customers audit the rule rather than discover it on an invoice.

4. Expose forecasting tools for the shared balance

Because one balance now funds four primitives (plus the legacy Residential Proxies add-on), a spike in proxy GB silently starves the request budget. A dashboard projection — “at current burn you’ll exhaust balance on day 18” with per-primitive attribution — would convert the shared balance from a footgun into a feature, and reduce the under-sizing that drives mid-cycle top-ups.

Key takeaways

1. Bill for outcomes you control, not effort you don’t. ZenRows’ success-only metering shows how to turn a reliability claim into a pricing guarantee — the customer never pays for a block, so price and product quality pull the same direction.
2. A shared balance lowers the buying decision but raises the forecasting risk. Pooling spend across products is a conversion win, yet it makes plans easy to under-size; pair it with forecasting tools or it generates support load and mid-cycle top-ups.
3. Price hardness with multipliers, not SKUs. A ×1/×5/×10/×25 ladder over one base unit keeps a pricing page legible while still charging more for expensive work — far cleaner than enumerating every proxy-and-rendering combination.
4. Publish derived rates before buyers have to compute them. ZenRows illustrates both sides of this lesson within its own timeline: its mid-2026 pricing hid the effective $/GB residential-proxy rate behind plan allowances, then the 2026-08-04 rebuild published a flat, explicit 25,000-credits/GB rate for the same product — closing exactly the credibility gap that made buyers suspicious in the first place.
5. Publish as far up the curve as you can. Self-serve pricing to $999/mo (down from $2,999/mo before the 2026-08-04 rebuild) still captures the mid-market buyers who bounce off “Contact sales,” leaving the sales motion for genuinely custom usage — the specific ceiling moved, but the strategic logic of pushing it as high as sales capacity allows didn’t.

UBP implications

1. Success-gated metering is a credible substitute for outcome-based pricing in infrastructure. Where a true business-outcome meter is impractical, billing only for technical success (“the scrape returned valid data”) gives buyers much of the same risk transfer — a pattern worth borrowing wherever a vendor’s value is “it worked.”
2. Difficulty multipliers are an underused middle layer between flat and fully metered pricing. They let a vendor charge proportionally to cost-to-serve without exposing raw infrastructure SKUs, preserving both margin and pricing-page simplicity.
3. Shared-balance wallets shift the UBP design problem from quota-setting to spend-forecasting. As more vendors pool usage across products into one balance, the differentiating capability becomes in-product forecasting and attribution, not the metering unit itself — ZenRows’ own two-month swap from a dollar balance (June 2026) to a credits balance (August 2026) shows how fungible the underlying currency actually is once the pooling is already in place.

Sources

• ZenRows pricing page — Free/Build/Launch/Growth/Scale tier prices, credit multipliers, Monthly/Annual toggle, and FAQ (accessed 2026-08-04)
• ZenRows docs: How pricing works — shared credit-balance model, success-only billing, cost multipliers, per-GB and per-minute mechanics, top-up and upgrade rules (accessed 2026-08-04)
• ZenRows pricing page — June 2021 archive — launch-era Starter/Growth/Scale request quotas and $11/GB residential slider (accessed 2026-06-04)
• ZenRows pricing page — December 2021 archive — Developer/Startup/Business credit repackaging (accessed 2026-06-04)
• ZenRows pricing page — March 2023 archive — comparison-grid redesign and $999+/mo Enterprise (accessed 2026-06-04)

Bottom line

ZenRows packages four primitives — Fetch, Extract, Batch, and Browser Sessions, plus a legacy standalone Residential Proxies add-on — into one shared-balance subscription whose defining move is billing only for successful results, a reliability claim turned into a pricing guarantee. Twice in 2026 it re-denominated that balance: a mid-year shift to explicit dollar CPM and per-GB metering was replaced within two months, on 2026-08-04, by a shared “credits” currency and a Free/Build/Launch/Growth/Scale tier ladder that dropped the cheapest paid entry from $69 to $19/mo. The model is honest at the headline and granular underneath: difficulty multipliers (1x/5x/10x/25x) mean the real bill is driven by your page mix, not your plan’s sticker price, so protected-heavy workloads can quietly require a far larger tier. For buyers who scrape mostly basic pages it remains among the simplest infrastructure purchases available; for those hitting anti-bot walls at scale, the ×25 multiplier is still the number that matters, now expressed in credits rather than dollars. Compare it against the rest of the pricing blueprint corpus to see how success-gated metering stacks up against seat-based and pure-usage peers.

Pricing timeline : Major events on a vertical axis

Each milestone below corresponds to a public pricing change, product launch, or material adjustment. Major events use a filled marker; minor adjustments use a faded one.

Credits currency replaces dollar CPM/per-GB; Free/Build/Launch/Growth/Scale tiers

Aug 2026

ZenRows rebuilt pricing around a shared 'credits' balance across four renamed primitives (Fetch, Extract, Batch, Browser Sessions), demoting Residential Proxies to a legacy add-on. The Developer/Startup/Business/Business-tab structure was replaced by Free/Build/Launch/Growth/Scale, each with 3 price/credit rungs; Free became a permanent $0/mo plan (5,000 credits) instead of a 14-day $1-usage trial, cheapest paid entry dropped from $69 to $19/mo, and a new Monthly/Annual toggle (17% off) replaced the old 3/6/12-month discount ladder. Enterprise now starts above 12.5M credits/month.

Shared-balance model: dollar balance, CPM, and Business high-volume family

Jun 2026

Regular tiers raised to Developer $69 / Startup $129 / Business $299; a Business tab adds high-volume plans Business 500 $499 through Business 3K $2,999, then custom Enterprise. One shared dollar balance now spans the Universal Scraper API (priced as CPM per 1,000 successful results), Scraping Browser (per GB + $0.09/session-hour), and Residential Proxies (per GB). Pay only for success; 404/410 count as success.

Comparison-grid pricing; Enterprise quoted at $999+/mo

Mar 2023

Same Developer $49 / Startup $99 / Business $249 headline, restyled as a feature-comparison grid with Enterprise shown as $999+/mo and per-tier uptime guarantees (99.1%–99.9%). A banner announced a ~1.1M euro funding round. Source: web.archive.org snapshot 20230315.

Repackaged to Developer / Startup / Business at $49 / $99 / $249

Dec 2021

Plans renamed and re-priced to Developer $49 / Startup $99 / Business $249 with an API-credit allowance (250K / 1M / 3M), and the standalone $11/GB residential slider was dropped in favour of 'Unlimited Bandwidth'. JS rendering ×5 / premium proxies ×10 / both ×25 multipliers became the explicit credit-cost model. Source: web.archive.org snapshot 20211205.

Launch-era request-quota plans (Starter / Growth / Scale)

Jun 2021

Earliest archived pricing: Starter $49 / Growth $499 / Scale $999 per month, plus a quoted Enterprise tier, sold on a flat monthly request quota (50K / 1M / 2.5M requests). A separate residential-bandwidth slider quoted $11/GB, and a free tier offered 1,000 requests. Source: web.archive.org snapshot 20210622.