Your Finance Team Deserves Better Than 5 Different Apps
One login. Total financial control.
Every extra finance tool adds more manual work, more approvals, and more opportunities for mistakes. What if everything lived in one place?
Ramp automates spending, expense tracking, payments, and procurement with AI, helping finance teams save time, reduce costs, and make faster decisions.
Every once in a while, a new product arrives that feels both magical and obviously incomplete. The people who notice the incomplete part early tend to do well.
Saeed Ezzati is a software engineer and solo founder in San Francisco. He spent about ten years as a full-stack engineer across large corporations and startups, always keeping a side project running on nights and weekends. When ChatGPT launched, he used it constantly - and within days, he kept hitting the same wall. The conversations were valuable, but the workspace around them was missing. Threads got buried. Search was limited. Exporting was awkward. Prompts were impossible to reuse.
So he built a fix for himself. Within a few days, he shipped the first version of Superpower ChatGPT to the Chrome Web Store - inside the first week of ChatGPT's public launch.
Today the extension has over 420,000 downloads, more than 150,000 weekly active users, and thousands of reviews across browser stores. Alongside it, his newsletter Superpower Daily has grown past 350,000 subscribers. Together, the ecosystem generates five-figure MRR - and Saeed still runs all of it alone.
Here is how it happened, and why the decision to stay free for nine months was the most strategic move in the whole story.
Shipping Inside the Window
The first version took two or three days to build. No funding, no team, no launch budget, no process. A laptop and a problem he personally wanted solved.
It was rough, and it was nothing like the product that exists today. But it solved one problem clearly enough that people immediately understood why it existed - and in a brand-new market, that clarity mattered more than completeness. Saeed is direct about the counterfactual: if he had waited to build the full vision, he probably would have missed the timing entirely.
What started as a simple extension grew into a productivity layer on top of ChatGPT: local history search and sync, folders, conversation export, message pinning, prompt management, writing style and tone options, and more. The roadmap came directly from user emails, reviews, Reddit comments, and Discord conversations - repeated pain he kept seeing in public, not features he dreamed up in isolation.
Free Was Not Charity. It Was Distribution.
For its first nine months, Superpower ChatGPT was completely free. That was a deliberate strategic decision, not an oversight or a delay.
Saeed's reasoning: in a new category, asking people to trust a paid product too early slows everything down. What he wanted first was users - real usage, real feedback, real reviews, real trust. Once the extension became part of people's daily ChatGPT workflow, a Pro tier could be a natural upgrade rather than a hard sell.
That created a second question: how do you support a free product with a fast-growing user base without ruining the experience with random ads? His answer became the second business. Because extension users were already interested in AI, he started sharing AI news and useful tools inside the product itself. Users liked it enough that it grew into its own standalone newsletter - Superpower Daily.
"Free can be a distribution strategy, a feedback strategy, and a trust-building strategy. You still need to know your revenue source."
The Hardest Part Was the Platform Underneath
The extension itself is deliberately simple: vanilla JavaScript, HTML, and CSS. No frontend framework. For browser extensions, Saeed prefers direct control over performance, bundle size, and how the code interacts with the page. Server-side components handle subscriptions, account state, newsletter workflows, and internal tooling.
But the engineering challenge was never the first version. It was survival. A normal web app controls its own environment. A browser extension lives inside someone else's product and has to keep working as that product evolves - and ChatGPT evolved constantly. Buttons moved. DOM structures changed. New features shipped. Old flows broke. Rate limits shifted. The simplicity of the stack turned out to be the thing that let him adapt fast enough to keep up.
His honest retrospective: if starting over, he would invest earlier in internal tooling, automated testing, and a more robust way to handle ChatGPT UI changes. He would also build more of the web-app side sooner, so the business depended less on browser extension constraints.
Two Revenue Streams, High Margins, One Person
The business runs on two engines:
Freemium subscriptions: The free plan carries most of the core value, with limits on advanced features. Pro removes the limits. Because the free product already had a large audience that understood its value, the paid plan sold itself - many upgraders had been daily users for months before paying.
Newsletter sponsorships: Superpower Daily became a serious business in its own right. Sponsors want to reach people who actively use AI tools, and 350,000 subscribers who arrived through an AI productivity product are precisely that audience.
Costs were almost nothing at the start - the main investment was time - and margins remain high because Saeed is still solo. No employees. The main costs are servers, software tools, and the hours required to support and maintain everything. The tradeoff he names openly: everything depends on his focus, which is why he has learned to prioritize repeated pain over loud feedback.
Product and Distribution Are the Same Job
Saeed's biggest growth lever was being early with something genuinely useful at the exact moment people were actively searching for ways to improve ChatGPT. He launched on Product Hunt, posted on Reddit, shared updates on X, and made the extension easy to find in browser stores. But he is clear that timing only worked because the product solved a real problem.
Reddit was especially effective early. People were already discussing ChatGPT workflows, prompts, and missing features. He joined those conversations with something closer to "I built this because I needed it - does it help you too?" than a marketing campaign. The honesty of that framing was the point.
His larger lesson is structural: product and distribution should not be treated as separate jobs. For Superpower, the product created distribution because users shared it. The extension created the newsletter audience. The newsletter brought people back to the product. Feedback from Discord, reviews, support emails, GitHub, Reddit, and social media all fed the same loop. The channels worked together because they were designed as one system.
The newsletter carries one more quiet advantage: it gives him a direct relationship with his audience instead of depending entirely on browser store discovery or social platforms. Writing about AI every day also forces him to stay close to the market - he notices trends, tools, and shifts in user behavior earlier than he otherwise would.
Five Things He Would Tell Someone Starting Out
Start with a problem you understand personally. You build a much more useful product when you are also a user. You notice small annoyances outsiders miss, and those annoyances often become the most valuable features.
Ship earlier than feels comfortable. The first version does not need to express the whole vision. It needs to solve one problem clearly enough for someone to care. Superpower would not exist at this scale if he had waited for the complete version.
Give people something genuinely useful for free — but know why. Especially in a new market where trust must be earned. Free is a distribution strategy, a feedback strategy, and a trust-building strategy. It is not the absence of a business model, and confusing the two is where free products die.
Build distribution while you build the product. Post in communities, write, share what you are learning, start a newsletter, build in public — whatever channel fits. A good product with no distribution is easy to ignore.
Listen to users, but do not blindly follow every request. Look for repeated patterns. The best ideas show up more than once, from different users, in slightly different words.
The Problem Keeps Getting Bigger
Superpower started as a small fix for one person's ChatGPT workflow. The reason it kept growing, in Saeed's view, is that the underlying problem kept growing with it. More people use AI every day, and their workflows are still messy: conversations disappear, useful answers get buried, prompts get rewritten from scratch, and finished work is hard to reuse.
His main goal is to keep making Superpower ChatGPT the best productivity layer for serious ChatGPT users - more organized, more searchable, more reusable, more personal. Alongside it, he wants to keep growing Superpower Daily, because the AI space is noisy and people need a trusted filter for what actually matters.
Longer term, the plan is more products around AI workflows - but only ones that save people time and reduce friction in work they already do. He is explicit about not building things because they are trendy. The filter is usefulness, and it has served him well so far.
The best place to learn more is the Superpower ChatGPT website. You can read Superpower Daily directly, and follow Saeed on X and LinkedIn.
Nobody wants to say no to their team. Ramp does it for you with AI-powered cards that automatically block out-of-policy spend. Let Ramp be the bad guy.
