Bhanu Teja, who sold Feather for $250,000 and focused on SiteGPT

Bhanu Teja P. is an Indian developer who created Feather, for running a blog with Notion, and SiteGPT, an AI chatbot for websites. What matters in his history is not that he sold Feather and then created SiteGPT, but the order: he launched SiteGPT first, and when it became difficult to grow both products together, he sold Feather. It is a case of turning a feature discovered in an existing product into an independent business and choosing which of the two businesses to focus on.

Turning the inconvenience of writing into a paid product

He quit his job as a software developer before making revenue from independent products. After leaving, he tried an education platform with another person for about a year but did not achieve results; after that, he worked as a freelancer for several months and saved living expenses that would let him last about two years, he explained. He did not leave with a successful product from the start; rather, through failures and contract work, he secured time to develop his own product.

Feather's predecessor was MDX.one, launched in June 2021. It was a service that published content written in Notion as a blog, but its operational structure caused problems, and the monthly server bill was charged up to about $10,000. He stopped new signups and rebuilt the product; when the rebuilt result was substantially different from the previous product, he decided to change the name too.

In this process, he revised both the technical architecture and the pricing model. While moving hosting to Cloudflare, he eliminated the free plan, making users who used it pay. He explained that even when he acquired many free users, he could not properly create paid conversion, and only support work and operating costs increased; this experience was the background for these changes.

The new product Feather launched on May 18, 2022. The point was not to make users learn a new writing tool, but to let them keep writing in Notion while running a search-engine-friendly blog. It was a structure that left Notion as the content management space, while the product handled separate website publishing and operation.

When acquiring early customers, he also did direct manual work. When WordPress users who showed interest on Twitter hesitated to adopt because of the work of moving their existing posts, he offered to migrate them for free. After spending hours copying posts into Notion and even building a blog to show them, he converted them into paying customers; in this case, what blocked the sale was not a lack of features but the burden of migration.

The metrics he published on March 6, 2023 were 191 paying customers and $4,180 in monthly recurring revenue (MRR). This was achieved about nine and a half months after launch, meaning that annualized to 12 months, annual recurring revenue (ARR) was about $50,000—not that he had already earned $50,000 over a year. At the time, he cited Twitter, word of mouth, and Google search as the main customer acquisition channels.

Feather's Add-On Becomes SiteGPT

SiteGPT's starting point was also a Feather customer. He came up with the idea of attaching an AI chatbot to existing customers' blogs so that visitors could explore the content by asking questions instead of searching for and reading posts themselves. While developing it, he judged that the same functionality could be offered not only to Notion-based blogs but also to general websites, and he split it off as a separate product rather than a Feather add-on.

The first version's functionality was to create a chatbot that, when a website address was entered, answered based on that site's content. In an October 2023 interview, he recalled building this minimal version in about 2–3 weeks, reusing the Cloudflare and database foundation he had used at Feather. Rather than developing a proprietary AI model, he combined existing AI API and web service components, and because he had to cover unexpected usage costs, he launched it as a paid product from the start.

The public launch took place on Twitter on March 30, 2023. At the time he had about 10,000 followers and had already been consistently sharing his product development process. After the launch post spread, it appeared on the Hacker News front page and ranked No. 1 for the day on Product Hunt, driving early traffic. Regardless of the short development period, an audience to whom he could show the product for the first time had already been built up.

SiteGPT reached $10,000 MRR one month after launch. Then, in an interview published in October 2023, he described MRR at about six months after launch as around $15,000. This is SiteGPT's own recurring revenue, which must be distinguished from Feather's performance or sale proceeds.

But not all early paying customers remained long-term customers. He said that the following month about $5,000 in recurring revenue was lost to churn, but new customer revenue of a similar size came in and offset it. Therefore, it is inaccurate to explain this point by saying that total MRR simply halved to $5,000. In his analysis, many early customers paid out of curiosity about a new AI product rather than an actual business need.

Revenue and profit were also different. In an August 2023 Mixergy interview, he described SiteGPT's monthly revenue as around $13,000–$15,000 and related monthly operating costs as around $4,000–$5,000, of which AI API costs were about $1,000. Because money also went into hosting and creating educational content, the disclosed revenue cannot be read directly as the founder's income.

The $250,000 Sale Was a Decision About Which Business to Focus On

As SiteGPT grew, it became difficult to spend enough time on both businesses at once. In a public Q&A at the time of the sale, he said he lacked the mental bandwidth to grow both products. Rather than because Feather failed, this can be seen as a decision to choose which of the two operating businesses to focus on.

The acquirer was Tibo, an entrepreneur he had known for some time. Tibo announced the acquisition of Feather on June 13, 2024, and on July 3 of the same year both sides disclosed that the final transfer had been completed. Bhanu also said that the full $250,000 had arrived in his bank account, confirming that it was transaction proceeds, not simply an asking price or offer amount for the sale.

At the time of the sale, Feather's MRR was roughly $6,000–$6,500 according to his own retrospective account, and he described the transaction price as about 3.5x ARR. The $250,000 at that time was the proceeds he received for handing over the business, a figure different from Feather's annual revenue or after-tax net profit. He also stated that monthly operating costs at the time were about $300, but this cost disclosure alone does not establish a net profit margin that also reflects the founder's labor.

The sale process exposed problems with account setup. Transferring access rights to the payment system was relatively easy, but because he had put the service on a personal Cloudflare account, moving the technical assets to the buyer's account took several weeks. He reflected that it would have been much easier if he had used a Feather-dedicated account from the start, showing that even a small SaaS, if it is to be sold, must be able to separate and transfer not only code but also operating accounts and infrastructure.

SiteGPT After the Sale and Changes in Operating Approach

It is also inaccurate to describe his businesses as fully solo operations across their entire lifespan. With Feather, he had help from a freelance designer, and for SiteGPT he initially developed and supported it alone before his brother Sai Dheeraj Pachipulusu quit his job and joined as cofounder. Dheeraj explained his reason for joining as the situation in which Bhanu alone had become overwhelmed handling both product development and customer support.

In a public post about two and a half years after launch, Bhanu disclosed SiteGPT's cumulative revenue of $500,000 and 135 companies using it. In the same post, he explained that he and Dheeraj, two people, were running it without outside investment. Here the $500,000 is cumulative revenue accumulated since the business began, and does not mean ARR or annual revenue.

However, the early rapid growth should not be expanded into sustained explosive growth over the entire life of the business. Even in December 2025, he disclosed crossing $15,000 in MRR, which is a level similar to the figure from the 2023 interview. Although the public materials from only these two points in time cannot reconstruct all the fluctuations in between, there is also no basis for stating that the growth rate immediately after launch continued.

The product expanded from its initial website-based Q&A into a tool for customer support work. The current official guidance includes adding knowledge through files and text, connecting to human agents, collecting lead information, analyzing conversations, and integrating with external business tools. It broadened its scope from simply generating answers to functions used in actual customer support processes.

The sales approach has not remained fixed as it was at launch either. It initially started without a free plan and without a free trial, but the current official site advertises a 7-day free trial. Therefore, what should be taken from his experience is not a principle to exclude free trials forever, but that he initially validated cost and willingness to pay and later adjusted the sales approach to match the product's state.

In a later retrospective, he also reflected on having made SiteGPT too general-purpose. A chatbot usable across all industries may not be sufficiently customized for any particular customer, and he stated that if he started again, he would deeply solve the problems of a specific industry and customer segment. The choice to broaden the market created an early opportunity, but he effectively judged that in the long term it is important to narrow the customer and use case.

In Bhanu's case, selling Feather is closer to an event that reallocated the developer's time and business assets than to retirement. The customer understanding and technical foundation gained from running the first product became the starting point for the second product, and when the two businesses began to divide his capabilities, he handed one over to another operator. It is a path for an independent developer that includes not only the ability to build multiple products but also the judgment to distinguish which products to keep and which to hand over.