Three Founders Built Supermeme.ai from a WhatsApp Group

Three People Who Met in a WhatsApp Group

In January 2022, Ramsri Goutham Golla started a WhatsApp group chat with Sanjeev NC and Nico Botha. The three had never met in person, and each was working on something different while they discussed what to build together. The goal was to launch a service that makes memes with artificial intelligence. Started without outside investment or an office, the project became Supermeme.ai.

The reason they joined forces was the work each had published online. Sanjeev was taking a no-code course while building a database that searched memes by emotion, and Ramsri was developing an AI meme generator while posting his progress on LinkedIn. The two looked at each other's work, talked, and decided to build together. Nico, who was connected to Ramsri through Twitter, joined them, and the three-person team was set.

Roles were split according to each person's experience. Ramsri took artificial intelligence and multilingual generation, Nico, a full-stack developer, took web application implementation, and Sanjeev, who had experience making memes and marketing products, took content and promotion. Most communication happened asynchronously on WhatsApp. The skills needed for AI development, service implementation, and customer acquisition were divided among the three co-founders.

The Steps One Meme Requires

Making a meme turned out to require several steps. After deciding on the story to convey, you had to find a suitable image format, then write a short line that fit the image's context. Finally you had to adjust the position and size of the text and touch up the image to suit where it would be posted. Even a familiar meme was hard to search for if you did not know its name, and after finding an image there was still the question of what to write to make it funny.

The three built an early product in about three weeks and released it in February 2022. When they posted the results on social media, interest gathered, but the developers themselves were not sure people would pay. The first paying customer appeared after launch, answering that question. In his first-year retrospective, Sanjeev thanked Nelson, the first paying customer.

Generating Lines, Searching Images, and Fixing Context

In the early product, a user picked a topic such as startups, sales, or programming and selected a template, the base image of the meme. The product then used GPT-3 to generate a line that fit the topic and the image. In an April 2022 product introduction, Ramsri also explained that templates could be searched by emotion or action. By connecting image selection and line writing, the product reduced the burden on users of having to memorize meme names or typical expressions.

Reading an image's meaning incorrectly threw off the result. A usage guide Ramsri published that same month featured an image of a plane suddenly reversing course, which the artificial intelligence described as an ordinary "flight route map." He let users edit the image description to focus on action and emotion, along the lines of "a person who suddenly changes a decision at the last moment." When the description was changed and the line regenerated, the result suited a startup's pivot or an investor's change of heart. The case showed how important it is to convey properly what situation an image expresses before generating a line.

Later the product removed even the step of choosing a template. In the "Text to Meme" feature, a user typed a sentence describing the situation to express, and the service suggested an image and a line together. The product took on more of the decisions a user needed to turn a story in their head into a meme. It was an improvement that changed the input method so a user could start just by describing the result they wanted.

Product Hunt, Multilingual Support, and the API

A few months after launch, the team introduced the product on Product Hunt and took first place for "product of the day." As visits and signups continued from outside English-speaking countries, the team invested in multilingual support earlier than expected. The range of supported languages grew to more than 110. By watching the users who actually arrived, the team expanded the languages and scope the service would handle.

In December 2022, the team released an API that let other programs use the meme generation feature. It was a feature users had requested often. Developers could connect Supermeme's features to their own software, and the team could welcome as customers the people building other products, in addition to those making memes directly on the website.

The Founders' Audiences and Search Traffic

The founders' existing audiences helped early promotion. The three had roughly 100,000 social media followers combined, and their audiences did not overlap much. As each introduced the product and its results, it could be shown to several groups at once. Sanjeev later looked back and said an important condition for early growth was the influence of the personal accounts they had built before launching.

As interest in generative artificial intelligence grew at the end of 2022, exposure went beyond the founders' audiences. Influencers began including Supermeme.ai in posts introducing "AI tools to watch." Meme generation was easy to understand from the output alone, and as users shared images they had made, the product was promoted along with them. In their first-year growth retrospective, the team named this sharing and outside introduction as a main source of traffic.

In February 2023, cumulative signups reached 100,000. Until then nothing had been spent on paid advertising, but the founders' content creation and personal account management handled promotion. People who saw the product on social media searched Google for "Supermeme," and traffic also came from searches for the use itself, such as "AI meme generator." People who knew the product name and people looking for a solution for the first time began arriving through different paths.

Once search traffic was confirmed, Sanjeev began spending more time on search engine optimization. He organized the homepage around search terms related to "AI meme" and published content steadily. According to the team's first-quarter 2023 report, the product ranked within the top three results for that group of search terms, and sometimes reached first place. A path for receiving search demand was added alongside waiting for social media posts to spread.

Paying Customers and Running It on the Side

In results announced in April 2023, the team disclosed 150,000 cumulative signups and 10 million memes generated. In the same post it said it had reached the milestone of 100 paying customers. While signups surged, the artificial intelligence usage bill also grew. To keep a product that many people try out, the team entered a stage where it had to cover generation costs while increasing paid use.

Around this time the team made its direction concrete toward a product marketing staff could use repeatedly. In a vision published in January 2023, it said it would reduce the monotony of marketing content and help people convey their thoughts through humor. It tried to reduce the effort of searching templates, writing lines, and editing images in real content production for blogs, presentations, and social media. The plan was to widen the product's use from a one-time experience of making a funny image to a tool used continuously at work.

Sanjeev later said in an interview that monthly recurring revenue was about 5,000 dollars and paying customers had passed 250. At that time cumulative signups had passed 500,000. Customers tried the service and paid without a separate sales call or product demo. It was a sales method that fit a team running a business with few people.

Even after revenue appeared, the founders kept doing other work. In the same interview, Sanjeev explained that, with no employees, the three co-founders ran the product part-time, and that he himself continued consulting and content production. They responded to important customer inquiries and feedback and improved the product little by little, without setting a specific deadline for revenue goals. The operating style they chose was fitted to securing the room to maintain their other activities.

Magic Memes and Pricing

Ramsri later introduced the "Magic memes" feature and widened the scope of generation to editing the image itself. Beyond finding a template that fits the situation entered and placing a line, it let users add related visual elements or change the template. It was an improvement that increased the work the product handles so that the sentence and image match better. The context problem that appeared early in the image description editing feature can be seen as extended toward handling the visual composition of the output.

In the product lineup as of September 2026, free accounts get 10 credits a month, and paid plans provide watermark removal, brand settings, and API access. Personal and marketing-team plans are divided by generation limits, the number of brands that can be managed, and API call volume. Customers pay for the features they need to make memes repeatedly or manage content for several brands. It is a structure in which a user checks generated results through a free trial and then subscribes according to their workload.

In Supermeme.ai's development, what stands out is that the team found the work to automate step by step. First it generated lines, then it let users find images by emotion, and later it arranged things so a user only had to write the situation and the product handled image selection and line placement. The point where Ramsri and his colleagues added value was in the process of reducing the decisions a user had to make to finish a single meme.