SiteGPT: 12 Questions Answered From the Starter Story Interview, With Timestamps

SiteGPT puts a chatbot on a business website that answers questions from the site's own content, and its founder grew it without paid marketing by building about 50 small free tools that rank in search. In the Starter Story interview, founder Bhanu reports $13,000 in monthly recurring revenue and around $500,000 in total revenue since launching in March 2023. Most of the site's search traffic, he says, lands on those free tools.

We transcribed the interview with transcribe.so and asked it the same questions we ask every episode. Each answer below is written in our own words and points to who said it and the exact second, so you can check it in the video. All figures are self-reported by the founder and we have not audited them. We are not affiliated with Starter Story, and we sell a transcription service.

What does SiteGPT do and how does it make money?

It gives a website an assistant that answers questions about the business's content around the clock (Founder, 01:47). You create a chatbot, add it to your site, and pay a monthly fee that depends on how many chatbots, how much content and how many messages you need (Founder, 02:37).

What numbers does the founder share?

$13,000 in MRR and about $500,000 in revenue since the March 2023 launch (Founder, 01:08). His funnel, per month: about 50,000 visitors, about 200 leads, about 60 trials, of which 25 to 40% become customers; average revenue per customer is about $100 across about 130 businesses (Founder, 02:53). Lifetime value per customer is about $1,700 to $1,800 (Founder, 03:20). More than a million people have visited the site since launch (Founder, 01:58).

Who is the founder?

A software engineer who joined a large startup after college, quit within eight months, and moved back in with his parents to build his own products; one of them reached around $6,000 in MRR (Founder, 03:54; 01:08). He later sold that product for $250,000 to focus on SiteGPT (Founder, 05:16).

"I'm a builder. I'm not a marketer." (Founder, 00:17)

Where did the idea come from?

In early 2023 he wanted to learn the new AI tools by building something for the 100 customers of his existing product: every one of them had a blog, so he built a chatbot that answers questions about it (Founder, 04:21). He then spun it out as its own product. The first proof of demand was fast: about $10,000 in MRR within the first month after launch (Founder, 04:49).

How long did it take to build, and with what?

Two weeks from idea to launch (Founder, 04:49). He builds mostly alone, as the host notes (Host, 11:50), with an AI coding assistant, a keyword research tool, his own chatbot for support, and standard tools for booking calls, analytics, email and docs (Founder, 11:56; 12:23).

How did it get from zero to $13,000 in MRR?

Without any paid marketing; everything came through organic channels (Founder, 01:58). In the same answer he says 60 to 70% of traffic comes from search, and almost 90% of that search traffic lands on the free tools. He has built almost 50 of them (Founder, 05:43).

"This is the cheapest and low-effort way to drive traffic to your site." (Founder, 00:34)

How do the free tools bring in search traffic, and how does each one lead people to the paid chatbot?

Each tool targets a search phrase and solves a small task related to SiteGPT, such as turning a PDF into markdown or naming a chatbot, so the people who use it are likely customers (Founder, 05:43). He reports about 50,000 clicks a month from search, almost 90% of them from these tools, and about 300 leads (Founder, 06:11). Under each tool sits a call to action that stretches the task into the product: if you can chat with one PDF, try chatting with all of your business content (Founder, 10:55). A new tool now takes him less than five minutes because an AI assistant copies the pattern of the existing ones (Founder, 06:11).

How does the founder pick the keyword each free tool targets?

Seven steps, shown on screen:

1. Open a keyword research tool and search with the box left blank, to see every keyword (Founder, 08:28).
2. Filter to keywords containing your terms, such as "AI" plus "generator" (Founder, 08:55). Broad ones are hopeless: one he shows has a difficulty of 87 (Founder, 09:20).
3. Keep keyword difficulty under 10 (Founder, 09:26).
4. Keep keywords with at least 1,000 monthly searches that relate to the product (Founder, 10:00).
5. List them with volume and difficulty in a planning doc (Founder, 10:28).
6. Write the call to action each tool will carry toward the main product (Founder, 10:28).
7. Score each one on volume, difficulty, effort to build and relevance, then build in that order (Founder, 10:55; 11:23).

How is it priced, and why?

A monthly fee scaled by chatbots, content and messages (Founder, 02:37), averaging about $100 per customer (Founder, 02:53). The interview gives no plan prices, no reason for the pricing and no price change.

What does it cost to run?

Marketing spend is zero (Founder, 01:58). He lists his tools but not what they cost (Founder, 11:56), and the interview does not discuss margins or staff.

What is the founder's advice?

"I would not spend months and months on launching something." (Founder, 12:48)

Launch with only the core feature and let user feedback decide where the product goes (Founder, 12:51). The host sums it up as launch first, then figure out what happens next (Host, 13:11). The interview does not describe a failure or mistake of his own.

What did the hosts take away?

The main host calls free tools one of his favorite tactics and says builders who hate marketing should market by building (Host, 13:25); earlier he noted that making the tools is actually fun for a builder (Host, 07:44). The second host agrees:

"This is almost like a little hack that lets you drive traffic and focus on what he likes to do." (Host, 13:49)

The main host also notes how AI cut the build time per tool to minutes, compared with his own free tools for Starter Story years ago (Host, 14:01; 14:13).

What to copy

• Let keyword data choose the tool. Low difficulty, real volume and relevance to the product decide what gets built, not what is fun to build.
• Give every free tool an exit. The call to action should turn the small task into the bigger one your product does.
• Template the tools. Once a few exist, each new one is a copy of the pattern, which keeps the cost per page tiny.

More from this series: free tools as an SEO engine, every Starter Story episode answered, and our free YouTube transcript tool if you want to run the same questions on any interview.