The pitch meeting lasted all of thirty seconds. Steve Hanov walked into a venture capital office ready to present his suite of software products, only to be stopped cold by a single question from across the table: "What do you even need funding for?"
For most founders, that would signal rejection. For Hanov, it confirmed he'd built something different.
Operating from Waterloo, Ontario—the same Canadian tech hub that gave birth to BlackBerry—Hanov runs what he describes as a portfolio of six software products, several of which he claims generate more than $10,000 in monthly recurring revenue, though these figures remain unverified by independent sources. His entire technical infrastructure costs him roughly $20 per month. There are no AWS bills climbing into four figures, no Kubernetes clusters humming away in data centers, no engineering team burning through payroll. Just one developer, a small collection of servers, and a technology stack that might elicit raised eyebrows in most Silicon Valley boardrooms.
Building in BlackBerry's Shadow
Hanov's story starts in an earlier era of tech. He spent nearly a decade at Research In Motion during what turned out to be the company's peak years, from 2003 to 2012. He worked his way up to Team Lead on EDGE/LTE firmware before landing in the Office of the CTO. Before RIM, there were stints at Microsoft and Corel. His academic credentials include two mathematics degrees from the University of Waterloo, earned between 1998 and 2009.
But during those corporate years, something else was happening. Side projects. Weekend experiments. Tools built to solve specific problems that Hanov or his peers encountered.
WebSequenceDiagrams arrived on March 25, 2010, while he was still at RIM. The concept was elegantly simple: type in text, receive a UML sequence diagram. It predated the widespread adoption of PlantUML and gained enough traction that Cisco eventually incorporated it into their educational materials. Not a bad outcome for something built in spare hours.
A Portfolio That Sprawls
Over the years that followed, Hanov assembled what amounts to a catalog of internet utility businesses. There's RhymeBrain, a rhyming dictionary built on AI-driven pronunciation analysis. Zwibbler, an embeddable JavaScript vector graphics library sold through perpetual licenses priced at $5,999 for the complete package. X Reply, a Chrome extension designed to help users compose responses on X (the platform formerly known as Twitter). And eh-trade, a stock research tool that's been in beta testing, which Hanov has actively promoted to early users on LinkedIn.
Each product targets a different niche, each generates revenue through different mechanisms—subscriptions here, one-time licenses there, credit packs elsewhere. It's diversification executed at a micro scale, a way of hedging against the inevitable decline or stagnation of any single offering.
In a post Hanov published back in 2014, he laid out the early economics. That year, he reported roughly $6,900 in monthly revenue, distributed across WebSequenceDiagrams subscriptions ($1,600), server sales ($1,600), RhymeBrain AdSense revenue ($1,700), and Zwibbler licensing and consulting work ($2,000). More recently, he's claimed that multiple products now individually exceed $10,000 in monthly recurring revenue—though these self-reported figures have not been independently verified.
The Infrastructure That Shouldn't Work (But Does)
What makes Hanov's operation particularly striking is how little it costs to run. One virtual private server from providers like Linode or DigitalOcean, somewhere between $5 and $10 a month. He deliberately sidesteps AWS, EKS, RDS—the usual pantheon of modern SaaS infrastructure. Everything runs on 1GB of RAM, occasionally supplemented with a swapfile when memory gets tight.
The backend is written in Go, deployed as static binaries via a straightforward scp command. No elaborate deployment pipelines, no containerization overhead. For the database layer, SQLite powers every product. Hanov enables Write-Ahead Logging and sets synchronous mode to NORMAL, a configuration he maintains delivers non-blocking reads and writes with surprising concurrency on NVMe storage.
For AI workloads, he made a single hardware investment: a used RTX 3090 GPU with 24GB of memory, purchased for an estimated $900. He runs vLLM for production inference, uses Ollama for iterating on prompts, and Transformer Lab for fine-tuning. User-facing AI features route through OpenRouter, which aggregates multiple LLM providers and offers automatic fallback.
Development assistance comes via GitHub Copilot in VS Code, which Hanov notes uses per-request pricing that keeps his monthly bills under $60—considerably less than alternatives like Cursor would demand.
It's a stack that resembles something from 2012 more than 2024, which might be precisely the point.
The Revenue Question Mark
The financial claims warrant scrutiny, perhaps more than Hanov's public posts provide. While he states he's running "multiple $10K MRR companies," the revenue model details tell a more complicated story. Zwibbler's pricing is built around perpetual licensing, not monthly subscriptions. X Reply sells credit packs at one-time prices of $9, $19, and $49—again, not recurring revenue in the traditional sense. WebSequenceDiagrams offers both a SaaS version and a server edition, but current pricing specifics weren't readily accessible during recent research.
What does hold up to scrutiny is longevity. WebSequenceDiagrams has been operational for more than fifteen years. RhymeBrain remains active and appears in third-party tools. Zwibbler shows up in real-world implementations. These aren't flash-in-the-pan experiments but sustained businesses, even if the exact revenue breakdown remains somewhat opaque.
When Distribution Becomes the Bottleneck
In a post on Hacker News, Hanov acknowledged that distribution has emerged as his primary challenge. A decade ago, building the software itself was the hard part. Now, with AI assistance accelerating development cycles, attracting users matters more than anything else.
His recent LinkedIn activity reflects this evolution. He's discussed growing from 200 to 10,000 daily impressions while promoting X Reply. He's actively recruiting beta testers for eh-trade. The developer who once could build products and let audiences discover them organically now finds himself in the business of marketing.
There's something almost contrarian—anachronistic, even—about Hanov's approach in today's landscape. While peers raise tens of millions and burn through cloud infrastructure budgets that could fund municipal governments, he operates a portfolio of profitable software on the equivalent of a few streaming subscription fees. It works, apparently, because he's optimized for an entirely different objective: not explosive growth, but sustainable income. Not hockey-stick curves, but steady cash flow. Not a liquidity event five years out, but revenue this month and next.
The venture capitalist who asked what Hanov needed funding for might have understood something, though possibly not in the way he intended. When you've constructed products that generate revenue on costs this negligible, when you've mastered the particular craft of doing more with radically less, the traditional funding playbook begins to seem almost absurd. You're not building for an exit. You're building for a life that doesn't require one.
Whether that life scales—or whether it even needs to—remains an open question. But fifteen years in, Hanov's still here, still shipping, still profitable on $20 a month. In an industry obsessed with growth at all costs, that might be the most radical bet of all.
