Case studyAugust 28, 202613 min read

How Browserless turned open source into a $3.6M ARR funnel

hosted headless-browser automation API

browserless.io

$1.3M ARRSep 2023 revenue, still solo

$301K/mocurrent revenue on IndieHackers (~$3.6M ARR)

13,606GitHub stars

160M+Docker pulls

80%margins, self-serve, no paid ads

In short

A support headache at his day job became the product. Joel Griffith noticed he and his colleagues spent roughly half their time helping users get headless Chrome running, and only 5-10% of their time building real features — so he built Browserless to be the fix he wished he'd had.
Open source and blog posts did what paid ads couldn't. The core Docker image was free to try, and technical posts on Puppeteer/Playwright best practices brought in the first real users — meanwhile a paid-ads experiment ran for months and produced not a single sign-up.
Two slow years, then a decade of compounding. 100 users after nearly a year, 300 by mid-2019, 500 paying users and $28K/month by Sep 2019 — climbing to $1.3M ARR by Sep 2023 and a reported $301K/month (~$3.6M ARR) as of Aug 2026.
Still mostly solo, seven-plus years in. Joel ran Browserless alone (aside from occasional contractors) past the $1.3M ARR mark in 2023, only then hiring a first full-time marketer — margins held around 80% throughout.

The mechanism

None of this was a single lucky channel — it was a chain. Griffith's own experience of the support-time drain gave him a product worth building; open-sourcing it removed the trust barrier that normally forces a sales process; and once paid ads proved worthless, the content-plus-self-serve combination he'd already been running had to become the entire growth engine, and it held up well enough to carry the company solo past $1M ARR.

Living the pain first made the product obviously useful

Before Browserless, Joel Griffith and his colleagues were burning about half their working time just getting headless Chrome to run reliably for users, leaving almost no time for real feature work. Because he'd personally absorbed that support burden, he wasn't guessing at a market — he was building the tool he'd already proven he and others needed, which meant the first version could be aimed at a real, specific pain instead of a generic one.

my colleagues and I spent roughly half our time helping our users troubleshooting how to get headless chrome up and running. As little as 5-10% of our time was spent on new features and delivering business value.

Open-sourcing the core let developers self-qualify before they ever paid

Because the Docker image was free and open on GitHub from day one, developers could read the code and run it themselves before trusting Browserless with money — removing the trust gap that normally requires a sales conversation. That try-it-yourself proof, paired with technical blog posts answering exact Puppeteer/Playwright problems, meant the product and the content did the convincing that a sales team or ad budget would otherwise have had to do.

Being open-source... gives you a lot of time trying out the service and sees if it fits without spending a dime. It's something I value personally, and since I happen to also be a developer it gives me a glance at how our users feel as well.

When paid acquisition failed outright, content and self-serve pricing had to (and did) carry the whole funnel

A months-long paid-ads test produced zero sign-ups, which forced all acquisition weight onto the two channels that were already working quietly: blog content and word of mouth. Usage-based pricing ($0.00008/second, no ongoing fees) then removed the last friction point — a developer could try, use, and pay for exactly what they used without a sales call — which is why growth kept compounding on an 80%-margin, largely self-serve base for years without new hires.

After running ads for a few months I don't believe we got a single sign-up for it, so I pulled the plug.

How it went

The problem, the build, and an underwhelming launch

2017-11 → 2018-10

Joel Griffith started building Browserless in Nov 2017 (GitHub repo created 2017-11-17) after noticing that he and his colleagues spent roughly half their time helping users troubleshoot headless Chrome instead of building real features. He open-sourced the core Docker image from the start, letting developers try it before paying, but the launch itself brought no customers in the first week and a slow first stretch — it took nearly a year in production before Browserless crossed its first widely-cited milestone of 100 users.

Technical content and open source start compounding

2018-06 → 2019-04

By Jun 2018, Browserless had served 2 million total sessions. Griffith kept writing founder-authored technical posts on the docs blog aimed at exact developer pain points ("Puppeteer best practices", "Observations running 2 million headless browser sessions"), which drove consistent organic traffic. A months-long paid-ads experiment during this stretch produced zero sign-ups and was shut down. By Apr 2019 usage-based pricing ($0.00008/second, no ongoing fees) launched, and the platform hit 3 million sessions run in a single day.

300 to 500 paying users, still solo

2019-04 → 2019-11

Customers tripled from 100 to 300 in about 8 months (by Jun 2019), and by Sep 2019 — two years after starting — Browserless hit 500 paying users and roughly $28,000/month in revenue, with margins hovering around 80%. Griffith was still the only full-time person running the company; a SourceSort interview around the same time cited $24k/month. Growth was almost entirely self-serve, with Griffith posting build-in-public updates on IndieHackers and giving founder interviews (Starter Story, SourceSort, Brandfetch) to extend reach.

Three-year mark to $1.3M ARR, first hire, and the current 2026 state

2020-11 → 2026-08

By Nov 2020, three years in, Browserless was trending #5 on GitHub's Daily TypeScript repositories. By Sep 2023, still bootstrapped and largely solo, it had reached $1.3M ARR, 15,000 active users, and 9,000 GitHub stars — the point at which Griffith finally hired a first full-time marketer and invested in The Browser Conference and a proxy product. As of Aug 2026, Browserless's IndieHackers page shows $301K/month in revenue (roughly $3.6M ARR), with the company's About page citing 10,000+ paying customers, 13,606 GitHub stars, and 160M+ Docker pulls.

Milestones

2017-11GitHub repo created; Joel Griffith begins building Browserless solo, off a day-job pain point
2019-0928000~$28K/month revenue; 500 paying users after two years, still solo
2019-1124000~$24K/month cited in SourceSort interview on monetizing an open-source project
2023-09108300$1.3M ARR (~$108K/mo), 15,000 active users, 9,000 GitHub stars; hires first full-time marketer
2026-08301000$301K/month on IndieHackers (~$3.6M ARR); 10,000+ paying customers, 13,606 GitHub stars, 160M+ Docker pulls

Whether it fits you

Browserless's loop rests on structural conditions that won't transfer to every product. Run it if these hold for you; the open-source-plus-content, ads-don't-work approach has real costs that are easy to underestimate.

What it needs

You're selling to developers who will read your source code before they'll read your pricing page

The open-source trial funnel only works because the buyer is technical enough to check a GitHub repo and run a Docker image themselves. A less technical audience won't self-qualify the same way, and the trust-building mechanic collapses.

You (or someone on the team) can write genuinely useful technical content, not marketing copy

The blog posts that worked answered specific developer pain points (Puppeteer performance, session-handling at scale) because the founder had actually lived those problems. Generic content doesn't rank or convert the same way for a technical audience.

You can run lean and solo for years while the compounding takes hold

Griffith stayed the sole full-time person on Browserless from Nov 2017 through past the $1.3M ARR mark in Sep 2023 — this only works if the founder can absorb both building and support work personally for a long stretch without needing to hire to keep pace.

What it costs

Giving away the core product for free costs you some paying customers, permanently

Griffith explicitly accepted that open-sourcing the Docker image meant losing some monthly revenue from self-hosters, in exchange for the retention and trust benefit of users who'd already read the source and knew the product.

The first year-plus is genuinely underwhelming, with no shortcut

Launch produced no customers in the first week and a 'painfully slow' stretch afterward — Browserless didn't hit its first widely-cited milestone (100 users) until nearly a year after the repo was created. There's no evidence a faster path was available once ads had already failed.

Paid acquisition may simply not work for this audience, and you have to be willing to write it off entirely

Months of paid ads produced not one sign-up. Committing to this playbook means being willing to abandon a channel completely rather than optimizing it, and putting all of that budget and time into content and product instead.

The numbers we could verify

margin: Hovering around 80% (founder-stated, 2019); no later figure disclosed
pricing: Usage-based ($0.00008/second) plus dedicated plans, alongside a free open-source Docker image
churn: Not quantified. Founder described losing a top customer as an emotional low point but gave no churn rate
sales model: Almost entirely self-serve; higher-touch approach reserved for enterprise customers later on

Channels it actually used

Developer content SEO
Open-source (GitHub / Docker)
Word of mouth
IndieHackers / building in public
Self-serve usage-based pricing

Our read

The easy read is 'open source plus blogging worked' — but the sourced mechanism is narrower: Griffith built the exact tool he'd personally needed as an engineer, which is why his technical content rang true to other developers instead of reading as marketing. A team copying the open-source-and-blog playbook without that lived-pain grounding would likely produce content that doesn't convert the same way.

The failed paid-ads experiment is one of the more useful data points here precisely because it's a documented failure: it forced Browserless's entire acquisition weight onto content and word of mouth rather than letting paid spend become a crutch. That forced focus, not a deliberate top-down strategy, may be what let the content channel compound as hard as it did.

Nine years of a solo (or near-solo) founder running a company past $1M ARR is unusual enough that it's worth treating as the load-bearing variable in this story, not a footnote. The open-source-plus-content mechanism may be replicable by a team, but the specific margin and headcount discipline documented here (80% margins, one hire made only after $1.3M ARR) is a founder-specific choice as much as a channel choice.

Sources

saas · developer-content-seo · open-source-trial-funnel · self-serve-usage-pricing · solo-founder-bootstrap