The practical question behind this tool is one more builders should be asking: if people are increasingly turning to ChatGPT, Claude, or Gemini instead of Google to find experts, products, and companies, what does your presence actually look like inside those models? A new site, intheweights.com, gives you a concrete answer by querying multiple LLMs simultaneously and scoring how consistently they recognize you.
The mechanics are straightforward. You enter a name or entity, the tool fires queries at several frontier and smaller models in parallel, clusters the responses by similarity, and surfaces an aggregate recognition score. The clustering step matters — it filters out noise and hallucination artifacts, giving you a cleaner signal about whether the models genuinely have coherent knowledge of you versus producing vague or contradictory outputs.
Why does this matter now? Web traffic data increasingly shows referral visits from AI assistants growing while traditional search referrals plateau or decline in certain niches. If a model has weak or inaccurate representations of your work, that translates directly into missed discovery — the AI equivalent of ranking on page five of Google. Unlike SEO, however, the levers for influencing what ends up "in the weights" are less obvious and involve things like publication presence, citation patterns, and the quality of publicly indexed content.
For builders and professionals, the immediate use is a baseline audit. Run your name, your company, your product. Note which models recognize you well and which draw blanks or confuse you with someone else. That gap tells you where your public footprint is thin and which content investments might improve model-side representation over time — particularly on platforms and in formats that training pipelines tend to prioritize.
The tool was built in a few weeks as a side project, which shows in its current scope, but the underlying concept is genuinely useful infrastructure for anyone thinking seriously about AI-era discoverability. Expect this category of "LLM presence auditing" to grow into a proper discipline as the share of AI-mediated traffic continues to rise.