Pivoting an internal tool into a 7-figure business
Lukas Böhler built an internal tool for his agency and validated it with his clients. Then, he launched, rebranded, and grew Gleap to a 7-figure ARR.
Here's Lukas on how he did it. 👇
Bootstrapping a 7-figure business
I'm Lukas, cofounder and CEO of Gleap. My background is in software engineering — before Gleap, I cofounded and ran BoehlerBrothers, a software agency in Dornbirn, Austria, where we built apps and products for clients for years.
Gleap grew out of a problem we faced daily at the agency: Clients reported bugs in long, vague emails that developers couldn't act on. We built an internal tool to fix this, and validated it almost by accident — when clients saw it, they kept asking if they could use it too. That signaled this wasn't just our problem, so we launched it as BugBattle in 2020. In 2021, we rebranded to Gleap and spun it out of the agency as a standalone company with my cofounders Isabella Salzmann (COO) and Tobias Duelli (CTO).
Gleap has evolved from a bug-reporting tool into a full AI-powered customer support and feedback platform — our vision is software that heals itself. At its core is Kai, our AI agent suite.:
Kai handles tier-1 support conversations across channels.
Kai Resolve investigates issues using technical context.
Kai Code turns bug reports into ready-to-merge pull requests.
Kai PM clusters feedback and closes the loop with users.
Around Kai sits everything a modern software team needs to talk to its users: live chat, in-app bug reporting, a knowledge base, surveys, product tours, and a public roadmap — replacing a fragmented stack of tools, like Intercom, Zendesk, Instabug, and Canny — with one platform.
The progress speaks for itself: More than 4,500 software teams — from indie founders to companies like Microsoft, Squarespace, UNICEF, and Papa John's — use Gleap today, and our widget reaches roughly 250 million end users monthly.
We're bootstrapped and profitable, with about 75% of our customers in the US. Kai now resolves the majority of tier-1 support conversations for many of our customers before a human ever reads them.
We crossed 7-figure ARR and are profitable, growing 4-8% month-over-month, all organically.
Building the first version
We built the initial product as a side project inside our agency. We didn't set aside a budget or raise money for it — we carved out time between client projects.
That's the advantage of starting from an agency: we had the engineering talent, the infrastructure, and the cash flow already in place, so agency revenue and our own unpaid hours essentially funded the first version. Grants from the Austrian startup ecosystem later helped us bridge the gap when we spun Gleap out as its own company.
The first version was deliberately narrow: an SDK you could drop into your app that let users report bugs with a shake gesture or button tap, automatically capturing screenshots, console logs, network requests, and device data. Because we were our own first customer, the feedback loop was brutal and fast — our developers used it daily on real client projects, so we knew within weeks what worked and what didn't. Our agency clients became the first beta testers, and their reactions told us we had something worth pursuing.
We stood on the shoulders of the usual suspects: open-source frameworks for our SDKs and standard cloud infrastructure allowed a small team to ship native SDKs for iOS, Android, and the web far faster than would have been possible a decade earlier.
Overall, it took us roughly a year from the first internal version to the public launch as BugBattle — a handful of people built it, our own agency funded it, and we validated it on real projects before a single stranger ever paid for it.
A standard stack
Gleap is a TypeScript/Node.js shop at its core. The backend uses Node.js with MongoDB as its primary database, and as our data volume grew, we added ClickHouse for analytics and event data — the widget reaches around 250 million end users per month, so the write volume on events and sessions outgrew what made sense to keep in Mongo alone. We run a fairly standard stack around that: Paddle for billing, Postmark for transactional email, and New Relic for monitoring and observability.
The most distinctive part of our stack is the SDK layer. Because Gleap lives inside our customers' apps, we maintain seven native SDKs — JavaScript, iOS, Android, React Native, Flutter, Capacitor, C#, and Cordova — which means we must build and test every feature across all of them. That's one of our biggest ongoing engineering challenges: keeping seven codebases in feature parity with a small team, across platforms that each have their own quirks around screenshots, console capture, and network logging.
The stack changed most dramatically when we went AI-first. Building Kai, our AI agent suite, meant layering an entirely new stack on top of the existing platform: LLM orchestration, RAG pipelines over customer knowledge bases and technical context, and Langfuse for LLM observability and evaluation — because with AI answering the majority of tier-1 support conversations autonomously, we need to measure and trace answer quality the same way we'd monitor uptime. The hardest lesson was that shipping an AI feature is 20% prompting and 80% evaluation, guardrails, and context engineering.
A two-part business model
Gleap is a classic B2B SaaS subscription business with a usage-based AI layer. Our plans run from $49/month for solo developers up to $999+/month for enterprise customers with compliance needs (SOC 2, SSO, BYOK), billed annually or monthly with a 14-day free trial and no credit card required.
On top of the subscription, AI usage for Kai, our AI agent suite, is billed by token consumption and model choice. This two-part model is deliberate: The subscription covers the platform, and AI usage grows with the value customers derive.
We started charging almost from day one. Coming from an agency background, we never had the luxury (or the temptation) of burning VC money on free users — Gleap had to pay for itself, so validating willingness to pay validated the idea.
Our pricing has evolved significantly since then: We launched with cheap, per-seat plans around $19–119/month, and we eventually moved to value-based platform tiers with unlimited seats on higher plans. That shift mattered — per-seat pricing punished the exact behavior we wanted (whole teams living in Gleap), while platform pricing plus AI usage aligns our revenue with customer value instead.
Revenue expansion comes from three directions.
Plan upgrades: Teams start on Starter or Team and grow into Pro and Enterprise as they adopt more of the platform — Kai Resolve, Kai Code, and Kai PM are Pro-and-up features, giving customers a natural reason to upgrade as they automate more.
AI usage: As Kai handles a growing share of a customer's support volume, token-based revenue scales with it, without requiring renegotiation.
Consolidation: Because Gleap replaces a stack of point solutions — Intercom, Zendesk, Instabug, Canny — expansion often means customers move budget from three tools into one.
Going AI-first caused our biggest revenue inflection. Launching Kai changed both the deal size and the buyer conversation: We went from "nice-to-have feedback widget" to "this cuts your support workload in half," a budget line every SaaS company understands.
My advice for aspiring entrepreneurs: charge early, even if it feels uncomfortable — a customer who pays $19 teaches you more than a thousand free users. And revisit your pricing more often than feels reasonable. Most founders set a price at launch and revisit it once every two years; nearly every time we repriced based on value instead of habit, revenue grew, and churn didn't.
SEO and PLG
As I mentioned, our first users came from our existing network. As an agency, we deployed the product on our own client projects, and those clients — plus other agencies in our network — became the first paying customers. This gave us something most launches lack: Real production usage and testimonials from day one.
When we launched publicly as BugBattle, we did the usual startup launch circuit — Product Hunt, beta lists, developer communities — which brought the first wave of strangers and, more importantly, taught us that US software teams, not local businesses, were our real market. Today, around 75% of our customers are in the US, operating from Austria.
Content and SEO have compounded most for us. We publish consistently on topics our buyers actively search — AI customer support, support automation, tool comparisons against Intercom and Zendesk — and over the years, this has built an organic engine that brings in trial signups every day without paid spend. Alongside it, review platforms like G2 perform quiet but steady work: When a support lead shortlists tools, we appear in the comparison set with social proof.
The product itself is our most unusual channel. The Gleap widget sits inside our customers' apps and reaches roughly 250 million end users a month. A meaningful share of new signups comes from people who encountered Gleap inside someone else's product, either as an end user or as a team member of a customer. Every deployment is a small billboard in front of exactly the right audience: people who build software. We reinforce this with a self-serve motion — 14-day free trial, no credit card, setup in under an hour — so curiosity converts without a sales call, plus a startup program (50% off the first year) which gets us into companies early and grows with them.
The AI shift also changed our go-to-market strategy. With Kai, we stopped selling features and started selling an outcome — "the majority of your tier-1 support handled before a human reads it" — and that message opened doors a bug-reporting tool never could, including larger customers like Microsoft, Squarespace, and Papa John's, whose logos then sell themselves.
My growth advice: Pick the channel that compounds and be patient. Paid ads stop the moment you stop paying; content, reviews, and product-led distribution keep working while you sleep. And don't underestimate your existing distribution — our agency network gave us our first customers, and our own widget became our best-performing channel. Most founders have an unfair distribution advantage somewhere; the trick is to recognize it.
Advice for indie hackers
My advice?
Familiarize yourself with modern coding agents and tooling.
Build MVPs
Start charging early
Do customer support and listen to your users!
Never give up. Keep pushing every day — even a little improvement accumulates!
What's next?
We aim to be the platform that supports self-healing software development, allowing everyone to put their software on autopilot and focus on what matters to their business, so they don't lose resources on customer support, bug fixes, etc.
Leave a Comment