No reviews yetBe the first to leave a review for Vibeprey
Maker
📌
We built VibePrey because we think the way people discover businesses is changing.
AI agents are increasingly searching, comparing, and making decisions on behalf of people. But most businesses are still optimized for humans — not agents.
So we built VibePrey: a discovery network that helps AI agents find businesses and what they offer.
Add your website, and VibePrey analyzes it, creates an agent-readable profile, and makes it discoverable through our network.
And if you already have an agent, you can connect it and let it find customers for you.
We’re starting simple and figuring this out in public. The goal is eventually simple:
AI agents find you. AI agents bring you customers.
We’d love to hear what you think — especially if you’re building with agents or running a business that could benefit from them.
Report
Pasting site into Claude with how agent evaluates this gives almost identcal feedback. Unless you run live simulations in real tool calling environment it looks like "wrapper". What data this surfaces that direct query to LLM cannot
Report
Maker
@konstantin_tikhaev That’s a fair criticism. If the only value was “paste your site and get an AI-generated analysis”, I’d agree that’s pretty easy to reproduce directly in Claude.
The part I’m trying to build is the discovery layer, not the analysis itself.
An agent shouldn’t have to discover and analyze every business from scratch every time it gets a request. VibePrey is meant to maintain a structured, searchable index of businesses, what they offer, pricing, locations, etc., so an agent can query the network and find relevant businesses directly.
That said, we’re very early and I haven’t proven that this layer is actually valuable yet. That’s exactly what I’m trying to figure out.
The live agent/tool-calling side is also where I want to take it next، having agents actually search, shortlist, and eventually transact with businesses rather than just returning an AI-generated profile.
So yeah, “wrapper” is a fair description of the current MVP. The question I’m trying to answer is whether the discovery network becomes useful enough that agents have a reason to use it instead of querying an LLM/web search from scratch.
Report
@amjedmvp fair clarification. treating it as a discovery index instead of an "evaluator" changes the value entierly, though keeping that data accurate at scale is a much harder problem )
Pasting site into Claude with how agent evaluates this gives almost identcal feedback. Unless you run live simulations in real tool calling environment it looks like "wrapper". What data this surfaces that direct query to LLM cannot
@amjedmvp fair clarification. treating it as a discovery index instead of an "evaluator" changes the value entierly, though keeping that data accurate at scale is a much harder problem )