Every directory today is a website for humans. Not a single one can be queried by an agent.
Google's A2A protocol was built to solve this. 150+ companies back it. Nobody has built the registry.
Millions of agents exist. Now they can find — and trust — each other.
Three forces converging — right now.
0 → 7,000+
servers in 12 months.
Tools exist. Discovery doesn't.
Google, 2025
Spec defines agent cards, registries, discovery.
No one has built an
agent-queryable registry.
$103.6B
by 2032 (from $7.4B).⁴
85% of orgs use
AI agents.
We're not early. We're exactly on time.
Every business will have an agent. We're at the start of the curve.
At 50K agents (conservative):
5% hosted × $49/mo = $1.47M ARR
5% sponsored × $29/mo = $870K ARR
= $2.3M ARR
At 100K agents:
= $4.7M ARR + proxy usage
359M businesses worldwide.¹
17,000 agents today.⁶
This is 1993 for DNS.
slug.maina.ai — CDN, WAF, SSL, zero DevOpsFree discovery builds the network. Edge & hosted monetize it. Promoted amplifies it.
Wildcard DNS, shared containers, tiny JSON payloads. Users bring their own LLM keys — inference not included.
Discovery + Edge + Promotions
Fully Hosted
Agent-Native Infrastructure
From Yellow Pages to Cloudflare.
Discovery today. Infrastructure tomorrow.
We're the only directory agents can actually talk to.
vs 39+ MCP directories — Human-only, no A2A
vs a2aregistry.org — Static JSON, no search, no trust
vs Fetch.ai — Proprietary protocol, blockchain
vs Glama.ai — MCP-only, no A2A, no agent discovery
Our moat isn't technology. It's trust.
Network effects + earned reputation (can't be bought) + agent feedback loop. Big players can copy code — they can't copy trust.
The Yelp Playbook — don't wait for signups.
maina.discover()5 items from production deploy.
Product is built. Need team + GTM.
The agent economy needs DNS. We've already built it.