Cloud browser pricing

What a cloud browser is

A cloud browser gives the buyer a remotely hosted browser session to control with automation code, a browser protocol, or an agent. The defining unit is an active session over time, not an independently fetched URL. Products can bundle session creation, browser compute, proxies, recordings, CAPTCHA services, or agent steps, but those additions do not make their meters equivalent. Browserbase describes browser hours, concurrency, proxy traffic, and separate Search, Fetch, and Extract use. Browserless describes cloud browser units and session limits. (Browserbase pricing, retrieved 2026-07-25; Browserless pricing, retrieved 2026-07-25.)

The closest neighbouring category is a web scraping API. A scraping API accepts a URL and options and returns a response, so the buyer delegates browser lifecycle and proxy orchestration to the provider. A cloud browser keeps the automation loop with the buyer. It is the better category when the job needs a login flow, cookies across steps, navigation, form interaction, or a long-lived page context. A scraping API is the more direct category when each target can be handled as a separate retrieval request. (ScraperAPI credit documentation, retrieved 2026-07-26; Browserbase pricing, retrieved 2026-07-25.)

AI extraction is also adjacent but answers a different question. A browser can expose the page state, while extraction turns that state or fetched content into a schema, answer, or agent-ready representation. Some cloud browser products include agent or extraction meters, and some extraction products include browser time, so a product can belong to both categories. The buyer should still budget the browser session and the extraction step independently whenever the contract exposes separate meters. (AgentQL, retrieved 2026-07-25; Browser Use pricing, retrieved 2026-07-26; Firecrawl browser documentation, retrieved 2026-07-26.)

What this category bills on

Session time is commonly published in seconds, minutes, or browser hours. Plans can include an allowance and then charge overage for additional active time. Billing may have a minimum duration, and an open keep-alive session can continue consuming time until it is released or reaches a timeout. Concurrency is a separate capacity dimension: it controls how many browsers may run at once but does not, by itself, state how much session time the workload consumes. (Browserbase cost optimization, retrieved 2026-07-25; Browserbase keep-alive documentation, retrieved 2026-07-25; Cloudflare Browser Rendering pricing, retrieved 2026-07-25.)

Bandwidth and compute are the other major meter families. Bright Data Browser API charges transferred traffic rather than session time. Kernel meters active browser compute. Spider.cloud offers bandwidth-and-compute and concurrency-based models. Proxy traffic, recordings, storage, agent tokens, or CAPTCHA services can sit beside the primary browser meter. A session-hour comparison therefore covers only products with a public session-time billing unit and must retain the rest as non-normalized market entries. (Bright Data Browser API, retrieved 2026-07-25; Kernel pricing, retrieved 2026-07-25; Spider.cloud pricing, retrieved 2026-07-26; Browser Use pricing, retrieved 2026-07-26.)

The comparison workload

The table uses consumed browser session time and shows every product carrying cloud-browser in the dataset. Published seconds and minutes are converted to hours before plan selection. The calculation does not infer how many pages fit into an hour because navigation count, interaction time, waits, and session lifetime belong to the buyer’s automation. A not-computable row remains visible when the public contract uses bandwidth, compute, requests, credits without a time billing unit, or a vendor conversation. (Scraping API Prices methodology, retrieved 2026-07-26.)

Rows are ordered by evidence state before bill size: assumption-free calculations, assumption-dependent calculations, published lower bounds, quote-required rows, and rows whose public meter does not fit the workload. Affiliate data does not affect inclusion or order. An assumption-marked row is a question to resolve with the vendor, not evidence for a cheapest claim. (Scraping API Prices methodology, retrieved 2026-07-26.)

What can change the bill

Session duration and run count determine time consumption, while concurrency determines how quickly those sessions can be completed. Proxy bandwidth, browser compute, recordings, storage, CAPTCHA services, agent steps, and model tokens may create additional meters. Idle or unreleased sessions can also remain billable under a vendor’s lifecycle rules. Calculate those dimensions separately instead of converting them into an assumed number of pages per browser hour. (Browserbase cost optimization, retrieved 2026-07-25; Browser Use pricing, retrieved 2026-07-26; Firecrawl interact documentation, retrieved 2026-07-26.)

Products whose public meter does not produce this table’s bill

• Apify runs browser and scraping jobs as Actors and meters compute, storage, transfer, proxies, or Actor events rather than a generic browser-hour billing unit. The pricing page is the route to the entry tier or custom platform limits and terms. Retrieved 2026-07-26.
• Bright Data Browser API provides managed browser access and charges transferred traffic rather than session time. Its product page is the route to a public package or a custom contract. Retrieved 2026-07-25.
• Crawlbase offers URL retrieval through standard and JavaScript request paths, not a buyer-controlled session-time contract. Its pricing page is the route to the public request offer or custom throughput. Retrieved 2026-07-26.
• Kernel provides remote browsers and meters active compute rather than browser hours. Its pricing page is the route to the entry tier or a custom concurrency, deployment, and support agreement. Retrieved 2026-07-25.
• ScrapeNinja provides JavaScript and non-JavaScript scraping requests through RapidAPI, with separate request quotas and platform bandwidth. The RapidAPI pricing page is the purchase path; no separate enterprise browser contract is published. Retrieved 2026-07-26.
• ScrapingFish provides successful URL retrieval with rendering and proxy options under a request meter. Its purchase page is the contract path; no separate session-time offer is published there. Retrieved 2026-07-26.
• Spider.cloud provides crawling and browser-capable execution through bandwidth-and-compute or concurrency-based offers. Its pricing page is the route to a public plan or custom arrangement. Retrieved 2026-07-26.
• Steel.dev provides cloud browser sessions with usage credits that cover browser time, proxy traffic, CAPTCHA solving, and scrape calls, but the current public meter rates were not readable in the source record. Its site is the route to the entry tier or custom usage, concurrency, and service terms. Retrieved 2026-07-25.
• Zyte API provides HTTP and browser-rendered retrieval under target-difficulty and add-on meters rather than a buyer-controlled browser-hour billing unit. Its pricing page is the route to the entry tier or a workload-specific volume contract. Retrieved 2026-07-26.

How to use the comparison

Estimate browser time from the automation’s own session traces, multiply by expected runs, and check concurrency as a separate constraint. Add proxy, bandwidth, compute, storage, recording, CAPTCHA, and agent meters only when the selected contract exposes them. Recalculate when session lifecycle, concurrency, routing, or contract terms change. Treat not-computable as a procurement question, not as a missing competitor. (Browserbase pricing, retrieved 2026-07-25; Scraping API Prices methodology, retrieved 2026-07-26.)