---
title: "AGENYRA"
description: "Why a distribution layer for AI products has to start as structured data, and why the public repo only proves the site that says so."
canonical: "https://www.ayushrai.site/work/agenyra/"
last-updated: "2026-10-16"
---

# AGENYRA

Source: https://www.ayushrai.site/work/agenyra/
Markdown: https://www.ayushrai.site/work/agenyra.md

Why a distribution layer for AI products has to start as structured data, and why the public repo only proves the site that says so.

Status: in-progress. Type: product. 2026–present.

## Problem

People cannot reliably discover, understand, evaluate, and reach AI products. Marketing pages and keyword lists collapse task, inputs, deployment, and cost into adjectives.

## Context

AGENYRA is aimed at people with a task and at builders who should describe a product once. The public site, agenyra.space, went up with the repository on 29 Sep 2026. It is a waitlist and an explanation, not an open catalog.

## Role

I built the public Next.js site, the waitlist, and the stage model in AGENYRA_product. The site describes a private build for taxonomy, profiles, search, and builder onboarding. That build is not in the public repository, so this note does not walk through its code.

## Thinking

Discovery and distribution are different jobs. Search can run over structured fields before any product is placed inside another tool or agent. Intent-aware matching is a research question until a system does it. I kept the stage vocabulary in one module so a feature cannot quietly become 'launched'.

## Execution

The public app is Next.js 16, React 19, and Tailwind CSS v4 on Vercel. Waitlist rows go to Supabase through server-side access, with row-level security and no client key. The product model on the site is six stages: discover, understand, classify, match, distribute, and measure. Foundation work is marked building. Structured search is experimental. Distribution is exploring. Network capabilities are planned. Contextual intelligence is research.

## Outcome

Visitors can read the model and join a waitlist. They cannot search a public product graph, and the site says the private build is not open. No usage, customer, or revenue figures are part of this write-up. What works today: the public Next.js site, waitlist to Supabase, and an explicit stage vocabulary. What remains incomplete: a public product graph, intent-aware matching, and any claim that distribution is launched.

## Lessons

- A stage enum is a product decision. Without it, the homepage and the roadmap drift apart.
- Publishing the ambition early is useful only if the repository boundary stays visible: this repo is the site, not the matcher.

## Technologies

- Next.js
- React
- TypeScript
- Tailwind CSS
- Supabase
- Vercel

## Links

- Demo: https://agenyra.space/
- GitHub: https://github.com/ayushrai-hub/AGENYRA_product

## Needs confirmation

Unverified. Do not cite these as facts.

- Behavior of the private product graph is taken from the public site's own stage copy, not from code in AGENYRA_product.

_Unverified award metrics are omitted. Wording matches the public site._