Radius

Inspiration

Honestly, it hit us when we were just casually asking ChatGPT for restaurant recommendations. We realized - wait, nobody's Googling anymore. We're all just asking AI assistants for suggestions. And then it clicked: brands have no idea if they're even showing up in these conversations. Like, companies spend millions on SEO to rank on Google, but they're completely blind to whether Claude or ChatGPT is recommending them. That felt like a massive problem waiting to be solved, so we jumped in.

What it does

Radius helps brands figure out if AI actually knows they exist. We track how often brands get mentioned across ChatGPT, Claude, Perplexity, and Gemini - basically asking these AIs hundreds of relevant questions and seeing who gets recommended.

But we don't just count mentions. We look at six things: how often you show up, what the AI says about you (good or bad), whether you're positioned as a top choice, if the context makes sense, how you stack up against competitors, and whether the AI would actually recommend you. Then we give brands a clear score and show them exactly what's working and what's not.

How we built it

We basically built a system that talks to multiple AIs at once and makes sense of what they say. Used Next.js for the frontend because we needed something fast and responsive. The tricky part was getting all these different AI platforms to play nice together - that's where Model Context Protocol came in clutch for orchestrating everything.

Spent a ton of time on prompt engineering to get consistent, structured data back from these AIs. Built a scoring system that actually makes sense instead of just throwing out random numbers. And designed the whole interface to be super clean - we wanted marketing folks to actually understand what they're looking at without needing a PhD.

Challenges we ran into

Oh man, where do we start. Rate limits were brutal - you can't just spam APIs with thousands of requests. Had to build smart queuing and caching to work around that.

Every AI platform responds differently, which made standardizing scores really hard. ChatGPT might give you a paragraph, Claude might give you a list, Perplexity structures things totally differently. Getting them all to speak the same language took forever.

And honestly, the biggest challenge was making this simple. We could've built something super complex with a million metrics, but then nobody would use it. Figuring out how to show "here's what matters and here's what to do about it" took a lot of iterations.

Accomplishments that we're proud of

We actually got this working across four different AI platforms simultaneously, which felt impossible at first.

The scoring system is something we're really proud of - it's transparent, you can see exactly why you got the score you did. No black box BS.

Built something that non-technical people can actually use and understand. That was huge for us because we didn't want this to be another tool that only data scientists can figure out.

And just the fact that we went from idea to working product that solves a real problem brands are facing right now - that feels pretty good.

What we learned

Transparency matters way more than we thought. Brands don't just want a score, they want to know why.

Each AI has its own vibe like ChatGPT tends to be more conversational, Claude's more analytical, Perplexity focuses on sources. You can't optimize the same way for all of them.

Prompt engineering is honestly wild. Changing one word can completely change the response you get. It's part science, part intuition, and a lot of trial and error.

And we're early to this space, which is exciting. This is like SEO in 1998 the brands that figure out GEO now are going to dominate in a few years.

What's next for Radius

Short term, we want to add competitor tracking so you can see how you stack up against others in your space.

Then we're thinking about building automated recommendations like "hey, based on what we're seeing, you should probably update your About page to mention X more."

We want to help brands actually create better content for AI visibility, not just measure it. Like a content optimization tool specifically for GEO.

Long term, we're looking at enterprise features for bigger teams, expanding to voice assistants, and maybe even predictive analytics so brands can stay ahead of trends instead of just reacting.

The goal is to make Radius the go-to platform for any brand that wants to show up when it matters - in AI conversations.

Built With

• claude
• gemini
• javascript
• openai
• python
• replit
• traxcon