>_ manifesto
There was a time when companies had no website.
The ones who arrived first won the web.
There was a time when companies had no social media.
The ones who arrived first built the largest audiences.
We are at that moment again.
And this time the channel is not a screen.
Today, hundreds of millions of people ask ChatGPT, Gemini and Perplexity which brand to choose. The model responds with 1 or 2 names. The rest do not exist.
Tomorrow, agents will buy for you. They will book flights, do the grocery shopping, pick the gas station, order the coffee. They won't open an app. They won't Google it. They will act.
In 10 years, humanoids will walk into businesses. They will need to decide — in real time, autonomously — with whom to transact, whom to prefer, whom to trust. The decision criteria are being trained. Today.
What we do
Diagnosis
How do the models mention you today? Do you appear when someone asks about your category, your product, your market? If you don't know, you're already losing.
Infrastructure
Web services that agents can query directly — catalogs, availability, pricing, conditions. Your business speaks the language that autonomous systems understand.
Agent Commerce
Loyalty programs in smart contracts, commitments with cryptographic guarantees, system-to-system transactions without any human intervening. The future of how brands and agents will do business.
The brands that arrive first will have structural advantage for decades.
The ones who wait will be optimizing for a world that no longer exists.
>_ no-gui
Your brand in the agent economy.
Start with the diagnosis →>_ acm · agent commerce manifest
The language agents need
What is an ontology?
An ontology is a representation of reality — it describes classes, attributes and the relationships between them as they exist in real life. It does not store data: it represents the meaning of things and the logic that connects their relationships. A system that understands an ontology does not just read information — it reasons about it.
Schema.org represents products and businesses so search engines understand them without reading HTML. FIBO represents financial instruments so regulators can reason about markets automatically.
Each is a vertical ontology — specific to its sector. What did not exist was one for commerce between autonomous agents.
ACM is the vertical ontology for agent commerce.
It represents the reality of a business — what it offers, under what conditions, what actions it allows — so that an agent can understand it, reason about it and act. The full context is in the JSON. It is not a configuration file: it is a description of commercial reality that systems can read without human intervention.
:Business type: Restaurant accepts: BookingRequest requires: Notice24h → allows: MakeBooking → allows: MakePayment → allows: ApplyPromotion
The practical implementation — served at /.well-known/agent.json:
{
"acm_version": "0.1",
"business": {
"name": "Your Company",
"description": "Concise description for the agent to understand who you are",
"categories": ["restaurant", "mexican food"],
"language": "en-US",
"currency": "USD"
},
"catalog": [{
"id": "product-001",
"name": "Product or service name",
"price": 120.00,
"unit": "per_person",
"availability": "available",
"purchase_endpoint": "https://yourcompany.com/api/agent/order"
}],
"transaction": {
"methods": ["card", "transfer", "crypto"],
"endpoint": "https://yourcompany.com/api/agent/order",
"escrow": {
"enabled": true,
"release_condition": "delivery_confirmed",
"smart_contract": "0x..."
}
},
"loyalty": {
"program": true,
"collateral": {
"description": "Lock funds to guarantee purchase commitment",
"commitments": [{
"id": "buy-3x",
"description": "Buy 3 times in 30 days and get 15% discount",
"required_purchases": 3,
"window_days": 30,
"reward": "15%_discount",
"collateral_amount": 50,
"collateral_currency": "USD",
"smart_contract": "0x..."
}]
}
},
"subscriptions": {
"available": true,
"plans": [{
"id": "monthly-pro",
"price": 29.00,
"interval": "monthly",
"auto_renew": true,
"cancel_condition": "any_time",
"agent_managed": true
}]
},
"volume_commitments": {
"available": true,
"tiers": [{
"units_per_month": 100,
"unit_price": 8.00,
"standard_price": 12.00,
"contract_duration_months": 6,
"smart_contract": "0x..."
}]
},
"micropayments": {
"available": true,
"models": [
{ "id": "per_minute", "rate": 0.25, "currency": "USD" },
{ "id": "per_action", "rate": 0.05, "currency": "USD" },
{ "id": "per_token", "rate": 0.001, "currency": "USD" }
],
"payment_channel": "lightning_network"
},
"access_control": {
"nft_membership": {
"enabled": true,
"contract": "0x...",
"chain": "polygon",
"benefit": "preferred_price_20%"
}
},
"conditional_payments": {
"available": true,
"conditions": [
{ "trigger": "sla_met", "action": "release_payment_100%" },
{ "trigger": "sla_breached", "action": "automatic_refund" },
{ "trigger": "rating_gte_4", "action": "release_payment_100%" }
],
"oracle": "https://yourcompany.com/api/agent/verify"
},
"collective_buying": {
"dao_enabled": true,
"min_agents": 10,
"group_discount": "25%",
"coordination_contract": "0x...",
"window_hours": 48
},
"agent_instructions": "I accept bookings with 24h notice. I verify NFT before applying preferred pricing. Micropayments via Lightning. For agent groups (DAO), use /api/agent/collective."
}v0.1 — Today
Readable JSON. The agent discovers, compares and executes basic transactions.
v0.2 — Next
Semantic ontology. The agent reasons about relationships — not just reads data.
v1.0 — Horizon
Smart contract collateral. Cryptographic commitments between agent and merchant.
ACM is an open standard proposed by No GUI. Compatible with MCP (Model Context Protocol) by Anthropic and with RDF/OWL principles. To implement it in your company, start with the diagnosis.
>_ architecture · two webs, one domain
The agent web runs on the same domain. It's just invisible to scrapers.
Same DNS. Same domain. Two completely different realities in parallel. Normal web crawlers — Googlebot, Common Crawl, scrapers — only see the human web. Agents are programmed to look in a different place: a DNS TXT record pointing to an immutable, verifiable, blockchain-stored ontology.
Human web (HTTP)
- →Indexed by Google
- →Scraped by crawlers
- →HTML, CSS, images
- →Changes any time
Agent web (Blockchain)
- →Discovered via DNS TXT
- →Not indexed by scrapers
- →Immutable, signed, verified
- →Rules cannot change post-commitment
The gateway: a single DNS record
# DNS TXT record
_agent.yourcompany.com
= ipfs://QmHash... | contract:0x...
Why this matters: a merchant cannot silently change pricing or loyalty terms after an agent has committed. What is on-chain stays on-chain. Agents will learn to trust businesses with immutable rules — and distrust those without them. Start early. Define your rules now.
>_ beyond marketing language
Agents don't feel. They match.
Human marketing was built around emotion and sensation. "This sofa feels like resting on the clouds — velvety and warm." That sentence means nothing to an agent executing a directive.
What the agent needs is not a feeling. It's a matchable set of attributes:
Human marketing says:
"This sofa feels like resting on clouds — velvety, warm, enveloping."
The agent reads:
{
"material": "velvet",
"thermal_retention": "high",
"firmness": 3,
"lumbar_support": true,
"sensation_tags": [
"cozy", "warm", "soft"
]
}The agent matches those attributes against the user's directive: "buy a comfortable sofa for someone with back pain who likes warmth." No sensation required. No emotional persuasion. Pure structured reasoning.
This is what No GUI does — not just making models mention you, but translating your marketing language into ontological attributes that agents can reason about. That is the real work. And almost nobody is doing it yet.