How SiteGPT grows

Built almost 50 free tools, each related to the main product so its users are likely SiteGPT customers (e.g. PDF-to-markdown converter, chatbot name generator). Each tool has a CTA below it pitching the main product in context, e.g. for a chat-with-PDF tool: 'You have tried chatting with one PDF. What if you can create a chatbot with all of your business content? Try SiteGPT.' New tools are built in under five minutes by telling Cursor to look at the existing free tools and create a similar one for a given keyword.

While learning AI tools over a weekend in early 2023, started building a blog chatbot for Feather's ~100 customers (everyone has a blog). Realized it shouldn't be limited to Feather users and split it out as a standalone product, SiteGPT, going from idea to launch in two weeks.

SiteGPT's traction left no time for Feather, so he sold Feather (then ~$5-6K MRR) to someone better placed to improve it, and went all-in on SiteGPT.

1) In Ahrefs Keywords Explorer, search with the box left blank to get all keywords. 2) Add 'include' filters for your space and tool type (e.g. 'AI' + 'generator'; also try 'builder', 'creator'). 3) Keyword difficulty filter under 10 (e.g. 'AI image generator' is KD 87, too hard). 4) Volume filter of at least 1,000 monthly searches. Keep ones relevant to the product (e.g. 'AI reply generator'). 5) List keywords with global volume and KD in Notion. 6) Write the CTA each tool would carry to the main product. 7) Build a table: keyword, volume, KD, build effort, relevance to product; prioritize high volume, low KD, easy to build, highly relevant.