We Built Mentionedby.ai to Track How AI Models Answer Questions
I'm Nikin, one of the creators of MentionedBy.ai — a platform we built to track how your brand is being mentioned inside AI-generated answers across ChatGPT, Claude, Perplexity, Gemini, and other LLMs.
We didn’t start this as a marketing tool. It began as a frustration. We noticed our startup wasn’t showing up in ChatGPT responses, even for queries where we clearly should have. Meanwhile, bigger or SEO-heavy brands were dominating AI answers — regardless of actual relevance. It wasn’t Google rankings anymore. It was model completions.
That sparked the question: how do you monitor and optimize your brand's visibility across these black-box models?
What It Does: Scrapes and audits responses from multiple LLMs using structured prompts.
Benchmarks your brand against competitors on frequency, sentiment, answer placement, and hallucination risk.
Tracks changes over time — e.g., if Claude suddenly stops mentioning you in a category you dominated last week.
Scores your Answer Engine Optimization (AEO) visibility across AI tools.
Sends real-time alerts if your brand is dropped or misrepresented by a major model.
Under the Hood: We run scheduled and real-time queries across models using OpenRouter and native APIs.
Responses are embedded using OpenAI + Cohere to cluster topic relevance.
We match mentions semantically (not just string matches), then run sentiment and hallucination checks (based on source correlation vs. output).
All model responses, timestamps, and prompts are logged to a time-series DB, and anomalies are flagged using fine-tuned classification.
The UI is built with bootstrap and a wordpress as the CMS, with a dark minimal design — and the backend runs PHP for session/user handling + Python for prompt orchestration and ML pipelines.
Why We Opened This Up: Most companies don’t realize AI is already shaping perception — silently. If your brand isn’t in the model, it doesn’t exist to the user.
We’ve seen:
-Founders being misquoted by AI.
-Startups dropped in favor of outdated competitors.
This felt important enough to make public.
If you're curious, we'd love feedback on how we could expose more data (we're considering a public LLM search ranking explorer too).
Here’s the link: https://mentionedby.ai (we are yet to launch publicly, so there is a waitlist)
Happy to answer any questions about the technical stack, model comparisons, or hallucination detection methods.
— Nikin Founder @ Synapse AI Labs (Built in Sri Lanka. Global by default.)