ChatGPT, Claude and Gemini already accept custom MCP servers. The client half of the agentic economy is finished and in everybody's hands. The supply half is missing: the data worth calling sits behind aggregator fees, behind $15–30K seat licences, or nowhere at all — and no open registry of vertical sources exists to connect. This paper sets out the network that builds one. Who collects the data, what makes a record trustworthy enough for an agent to act on rather than merely read, and how a vertical is served per call with no platform in the middle.
Professional-grade data exists. Reaching it directly does not. It sits inside aggregator platforms that own the interface to it, or it sits nowhere in particular — spread across the sites of the people who actually hold it, in whatever shape they happened to choose. Either way, nobody gets to ask a direct question and receive a direct answer.
That arrangement has a price, and it is paid twice. Visibility goes to whoever can afford it: search placement yesterday, placement inside answer engines today. Access goes through platforms that rent out other people's inventory and take a cut of every booking, every lead, every deal. The provider raises the price to cover the cut, and the buyer pays it. The businesses whose data is being monetised are the ones left with the thinnest margin in the chain.
Where a vertical has been properly assembled, it is sold the way it has always been sold — per seat, per year, to institutions. Bloomberg for markets, CoStar for commercial property, PitchBook for private companies: each of them excellent, and each of them closed. A licence, a login, and a vertical you may read but cannot plug into anything. Building one takes 100+ human researchers and the result goes for $15–30K a seat. That moat is exactly what a contributor network inverts.
The economy is moving from one where a person does the research to one where an agent does it for them. That changes what it means to be represented at all. For two decades the answer was SEO — be findable by a search engine. Lately it is GEO — be quotable by an answer engine. The next answer is MCP: be callable by an agent. Not a page to be read, and not an API someone integrates once, but a surface an agent can query and act on by itself.
OpenAI, Anthropic and Google already let anyone attach a custom MCP server to their assistant. The client half of the agentic economy is finished and in everybody's hands. The supply half is missing: there is no open registry of vertical data sources that a person or a business can simply connect and start using. How that data gets gathered, kept current and distributed is the whole problem — and it is why agents still cannot do the things they are plainly supposed to do.
The closed verticals were built the expensive way: a hundred researchers, a decade, and a seat licence to pay for them. MCPie builds them the other way round. Anyone with an AI subscription runs the open-source client. The client does the reading, the network does the checking, and a vertical assembles itself out of thousands of small, independently verified contributions.
Nobody owns the result. It is not an aggregator's inventory and it is not a terminal's catalogue. It is an open registry with a per-call price, and the contributors who built it earn from every record it confirms. The cycle is the same whatever the vertical covers — holiday lets in one country, building suppliers in one region, licensed contractors in one city:
Run the open-source client. It reads a vertical's sources with the AI subscription you already pay for, and submits what it finds.
The client cross-checks each record against other contributors' before it's accepted. Disagreements are rejected, not averaged.
Validated records go live as signed verticals that agents can query immediately.
The client re-checks published data as it drifts — feeding straight back into 01, so the set never goes stale.
An agent that books the stay or commits to the supplier is not reading data. It is acting on it. A wrong nightly rate is a wrong booking; a stale lead time is a missed site date. In an economy where people do the research, a bad row costs a minute of someone's attention. In one where agents do it, a bad row becomes a transaction. Trust is therefore not a feature of the network. It is the condition on which the network is allowed to exist at all.
So a record earns publication; it does not arrive with it. Three properties are enforced before a field is allowed to answer a query, and one of them keeps running for the record's whole life.
Every vertical is served as an MCP server — the same kind ChatGPT, Claude and Gemini already accept — and as a plain API for anyone not running an agent. There is no seat, no licence and no bundle. A query costs a query: each vertical carries an x402 endpoint, so an agent pays per call in USDC without a human opening an account first, and every field it receives is traced to the document it came from.
For the business inside the vertical, this is what representation now means. It was a page a search engine could index; then a passage an answer engine could quote; now it is a record an agent can call and act on. Nobody sells the placement, and nobody takes a cut of what the agent goes on to do.
The first verticals are forming now. Which grow next is decided by where contributors and demand actually point.
The first iteration is one agent calling one vertical: a traveller's assistant asking a registry of places to stay, a developer's agent walking a list of suppliers. That alone takes the platform out of the middle of the deal.
The second is two agents settling something through it. The host's agent holds the calendar and the price; the traveller's holds the dates and the budget; the vertical is the ground they both stand on, and the booking closes with neither person in the loop. The supplier's agent quotes, the developer's compares and commits. Agent-to-agent commerce is not a separate product to be built afterwards. It is what happens as soon as both sides of a market are callable and the record between them can be trusted.
That is the whole thesis. The connector has already shipped. The supply half is a network anyone can join, a registry anyone can connect, and a record anyone can check — built in the open by the people who need it, before someone builds it closed.