interactgen started from one observation: nowadays no one searches on google, everyone asks chatgpt. and when the answers come from ai, showing up inside those answers is the new seo. that is what interactgen sells, and i had two months to take it from empty repo to live.

the face of the product is one number. i made a scoring engine that rates how visible a brand is in ai answers and ranks it against its competitors, so a company can see exactly where it stands and who is ahead.

behind that number is a pipeline that never sleeps. every day it asks the five major answer engines, chatgpt, gemini, copilot, grok, and perplexity, thousands of questions about a brand and its competitors, collects every answer, and turns them into visibility analysis. around 10,000 results per org, a million results a day overall, at a 0% failure rate.

measuring visibility was only half the product. the other half was moving it. i built re-indexing, a proprietary algorithm that pushes a brand up in ai answers. with it, sprinto went from rank 9 to rank 2 in a month.

then came the question every customer asked next: is ai actually visiting my site? so i developed an image tag based tracker that spots llm and bot visits on a website and folds them into visitor analytics. it handles 200k pings a day.

scale kept us honest. the heaviest api paths were taking 2.5 seconds, so i rewrote them around mongodb in-memory aggregations and tightened the indexes until they answered in 300 milliseconds. the nightly run took 6 hours, so i moved it from aws sqs to bullmq with concurrent workers on redis streams, and it now finishes in 40 minutes.

in three to four months, we scaled interactgen from an empty repo to around 100 orgs and $400k in revenue.

founding sde. remix, node.js, aws, llms.