Browserless rents you headless Chrome, Firefox and WebKit in the cloud. You point Puppeteer or Playwright at its endpoint, or call its REST APIs for content, screenshots and PDFs, and Browserless runs the browsers, the captcha solving and the proxies. It also sells a self hosted license, which is rare in this market.
The pricing page lists units, and units are where most estimates go wrong. The rates below were read from browserless.io/pricing, including the plan data behind the monthly and yearly toggle, on October 9, 2026, and the yearly rates rechecked on October 24, 2026. Prices are in USD. The page opens on yearly billing, so the first numbers you see are the discounted ones.
The five Browserless plans and what each includes
| Plan | Monthly billing | Yearly billing, per month | Units | Concurrent browsers | Overage per unit |
|---|---|---|---|---|---|
| Free | 0 | 0 | 1,000 | 2 | none, requests rejected |
| Prototyping | 35 USD | 25 USD | 20,000 | 5, or 10 yearly | 0.0020 USD |
| Starter | 200 USD | 140 USD | 180,000 | 30, or 40 yearly | 0.0017 USD |
| Scale | 500 USD | 350 USD | 500,000 | 80, or 100 yearly | 0.0015 USD |
| Enterprise | custom | custom | millions | hundreds to thousands | negotiated |
Maximum session length grows with the plan: 2 minutes on Free, 15 on Prototyping, 30 on Starter and 60 on Scale. Every plan, Free included, gets BrowserQL, automatic captcha solving and access to Browserless's residential and datacenter proxies. Paid plans add bring your own proxies and screen recording. Enterprise adds private deployments, a commercial self hosting license, SSO, user roles and a dedicated account manager, and the page lists SOC 2 Type II and an available HIPAA BAA.
What a Browserless unit is, and the three things that spend it
Browserless defines a unit as a block of up to 30 seconds of browser time per browser connection. A connection that runs longer spends another unit every 30 seconds, and a session reconnect counts as a new connection with a fresh unit. So far that sounds like one unit per page.
Two other meters draw from the same pool. Proxy traffic costs 6 units per MB through residential proxies and 2 units per MB through datacenter proxies. A successful captcha solve costs 10 units. A real browser downloads everything a page asks for, HTML, scripts, styles, fonts, images and tracking calls, and every megabyte of it travels through the proxy.
That is the detail the plan cards hide. A retail page that weighs 2 MB loaded in full through a residential proxy spends 1 unit of time and 12 units of bandwidth. The Prototyping plan's 20,000 units are about 1,500 such pages, not 20,000. None of the review sites we read in the search results for Browserless pricing does this conversion; they restate the plan cards.
Browserless cost per 1,000 pages
The table assumes one REST call or one fresh connection per page, under 30 seconds, so 1 unit of time each. Light means images, media and fonts blocked in your script, about 0.3 MB through the proxy. Full means the page loads as a visitor sees it, about 2 MB. Monthly billing at each plan's included rate. Put in your own page weight and the method holds.
| Way of fetching | Units per page | Prototyping | Starter | Scale |
|---|---|---|---|---|
| Browser, no proxy | 1 | 1.75 USD | 1.11 USD | 1.00 USD |
| Datacenter proxy, light page | 1.6 | 2.80 USD | 1.78 USD | 1.60 USD |
| Residential proxy, light page | 2.8 | 4.90 USD | 3.11 USD | 2.80 USD |
| Residential proxy, full page | 13 | 22.75 USD | 14.44 USD | 13.00 USD |
| Residential light page plus one captcha solve | 12.8 | 22.40 USD | 14.22 USD | 12.80 USD |
Yearly billing takes 30 percent off, so the Scale row for a light residential page drops from 2.80 to about 1.96 USD. Keeping one connection open and loading a page every 5 seconds cuts the time part to about a sixth of a unit per page, which matters only on the no proxy row. Once a proxy is on, bandwidth is most of the bill.
For comparison, one successful page on our plans is 0.98 USD per 1,000 on Starter, 0.60 on Growth and 0.33 on Scale, with rotating residential and datacenter proxies, browser rendering, retries and field extraction in that price, whatever the page weighs. Yearly billing halves those. A failed page is not billed.
Three Browserless billing rules worth reading twice
Overage is invoiced in chunks the size of your plan. Browserless bills overage whenever it reaches the dollar amount of your subscription, at a plan change, or at month end. On Prototyping that means a fresh 35 USD charge each time you run about 17,500 units over. The emails warning you arrive at most once an hour.
A plan change resets the cycle and charges the full new plan. The FAQ says changing plans resets your billing cycle to the moment of the change and charges the full amount of the plan you move to. If you upgrade from Starter to Scale on day 20 because a job overran, you pay for Scale in full that day.
Free stops, paid keeps going. The Free plan rejects requests once its 1,000 units are gone. Paid plans never stop; they bill overage. That is friendly for uptime and unfriendly for a crawler with a bug that loads every image on a site through a residential proxy overnight.
Real monthly budgets on Browserless and on flat per request pricing
Three workloads on sites that need residential proxies, light pages at 2.8 units each, on the cheapest Browserless plan that fits, monthly billing with overage at the published rates.
| Pages per month | Units needed | Browserless | Flat per request |
|---|---|---|---|
| 50,000 | 140,000 | 200 USD, Starter | 49 USD, Starter |
| 250,000 | 700,000 | about 800 USD, Scale plus 200,000 overage units | 149 USD, Growth |
| 1,000,000 | 2,800,000 | about 3,950 USD on Scale list rates, or an Enterprise quote | 499 USD, Scale |
With full page loads the Browserless column grows about four and a half times. At a million pages a month Browserless will quote Enterprise below list, so ask before you rule it out. The shape of the gap holds, though: billing by the megabyte punishes heavy pages, and a flat price per successful page does not care what the page weighs. The same pattern shows up in our Browserbase pricing breakdown, where proxy gigabytes are the line that grows.
Where Browserless is the right buy
Automation that acts, not just reads. Logging in to portals, clicking through multi step forms, downloading files, running Lighthouse audits, rendering PDFs. A scraping API returns a page; Browserless hands your code a full browser with session replays to debug what it saw. For that work it is the better tool and its unit price is fair.
Sites that do not block. With no proxy and a reused connection, a page costs a fraction of a unit, well under 1 USD per 1,000 pages on any paid plan. Internal dashboards, partner sites and documentation fall here, and so does any team that already pays for its own proxies and plugs them in.
Teams that must self host. Browserless sells a commercial license to run the same service in your own cloud or data center, with Docker and Kubernetes guides. For a regulated team that cannot send pages through a third party, that option decides the purchase. If you go that way you still own the servers, so it helps to have a tool that provisions servers and ships zero downtime deploys for the cluster that runs the browsers.
High concurrency at a small budget. Starter allows 30 concurrent browsers on monthly billing and 40 on yearly; our Starter allows 5. If your workload is a burst of short sessions rather than a steady stream of pages, that matters.
Where flat per request pricing wins
Reading protected pages at volume. Retail catalogs, travel fares, marketplaces and review sites sit behind Cloudflare, DataDome, Akamai or Imperva and need residential traffic on every request, plus the occasional captcha at 10 units. That is exactly where the Browserless bandwidth meter runs, and where one price per successful page comes out three to forty times cheaper on the numbers above. Our Cloudflare scraping API and DataDome bypass API pages explain how the retries work and why blocked attempts are not billed.
Crawling whole sites. Browserless has no crawler; your code discovers links and queues pages. Our website crawler API takes a seed URL, a depth and include or exclude patterns and returns every page as JSON, billed per successful page like any other request.
Budgets a finance team can sign. 149 USD for 250,000 pages is a number. A unit pool that drains faster every time a retailer adds a hero video is harder to approve. If you are pricing other vendors too, the same arithmetic for Scrapfly is in our pricing series, and every plan of ours is on the pricing page.
Questions buyers ask
How much does Browserless cost? Free is 0 with 1,000 units and 2 concurrent browsers. Prototyping is 35 USD a month with 20,000 units, Starter 200 with 180,000 and Scale 500 with 500,000. Yearly billing brings them to 25, 140 and 350 a month and adds concurrency. Enterprise is quoted per customer.
What is a unit in Browserless? A unit is up to 30 seconds of browser time on one connection. Longer sessions spend another unit every 30 seconds and a reconnect starts a new unit. Residential proxy traffic costs 6 units per MB, datacenter 2 units per MB, and each successful captcha solve 10 units.
Does Browserless have a free plan? Yes. Free gives 1,000 units a month, 2 concurrent browsers, 2 minute sessions, captcha solving and access to the proxies, with no card. Requests are rejected once the units run out, so it suits testing an integration, not running a scraper.
How much do Browserless proxies cost? They are paid in units, not gigabytes: 6 units per MB residential and 2 per MB datacenter. A gigabyte of residential traffic is 6,000 units, which works out to 6 USD on Scale's included rate and 10.50 USD on Prototyping. A gigabyte of datacenter traffic is 2,000 units, 2 to 3.50 USD.
Is Browserless cheaper than a scraping API? For unprotected sites and interactive automation, often yes. For protected pages at volume, usually no: a light page through a residential proxy is about 2.80 USD per 1,000 on Scale against 0.33 to 0.98 USD per 1,000 on our plans, before any captcha.
Is there a Browserless alternative for scraping? If the job is reading pages rather than driving a browser, a scraping API that returns rendered pages as JSON at a flat price per successful request is the usual alternative, and our Browserless alternative page sets the plans against each other at three budgets. If you need scripted browser control, the comparable hosted option is Browserbase, which bills browser hours and proxy gigabytes instead of units; our Browserbase vs Browserless pricing comparison runs both through the same workloads.