Agentic data collaboration

Reach meaning anywhere.
Apply context everywhere.

The marketplace for the agentic web. Content indexed at the vector level, queryable by any agent, buyable by any platform, governed by open standards.

01 / Who it's for

Four seats. One venue.
Aligned incentives.

Data collaboration only works when every party at the table wins. Vectorplatz aligns the open web's content layer with the buy side, the agent layer, and the model layer through one venue, one wire protocol, and one trust framework.

/01Sellers

Publishers and data providers

Sell context at the right granularity. URL-level licensing beats domain-level licensing for any portfolio with mixed value. One integration reaches agencies, agent platforms, and AI labs. ARTF receipts make every use of your inventory auditable.

/02Buyers

Agencies and brands

Buy context with the precision of audience targeting and none of the identity exposure. One catalog spans many contextual data providers. IAB 3.1 compatible. AdCP wire protocol means the same buy works across DSPs.

/03Buyers

Agents and platforms

ContentGraph runs as an MCP server. Your agent grounds reasoning on graded, classified, trust-receipted context. Trust receipts let the agent prove its reasoning chain back to the user.

/04Buyers

AI labs and model builders

License context at the URL level for retrieval and training. Provenance receipts on every record. Publisher contracts negotiated once at the collaboration layer instead of one publisher at a time.

02 / Why collaboration

The agentic web is forming.
The data layer underneath has to be different.

The open web is 4 billion URLs deep and growing. Domain-level classification flattens it. Cookie-based identity broke it. Audience targeting projects a three-dimensional buyer onto a two-dimensional plane and calls the shadow a person.

Agents need context graded at the unit they consume, which is the URL. Publishers need pricing at the unit they create, which is also the URL. Agencies need taxonomies they already use, which are mappings on top of the same vector. AI labs need provenance receipts on training data, which only an indexed substrate produces.

A broker sells lists. A collaboration layer aligns incentives across parties that need each other to win. Vectorplatz operates the second.

03 / Pillars

Three pillars hold the collaboration up.

01 · Pillar

Transparency

Every record in the catalog carries provenance. Every transaction generates an ARTF receipt. Every receipt is signed, timestamped, and verifiable by any party. Buyers see what they bought. Providers see what was queried. Agents see what was matched.

02 · Pillar

Activation

One integration reaches the buyers and the agents already operating at scale. AdCP for transactions. MCP for agent retrieval. PubMatic, Equativ, and the major DSPs for traditional buy paths. IAB 3.1 for taxonomy compatibility. The wire protocols are open. The integrations are pre-built.

03 · Pillar

Collaboration

Multi-party data exchange without exposing the underlying content or the underlying buyer. Publishers license access without surrendering control. Buyers query meaning without aggregating identity. AI labs train on graded inventory without scraping the web themselves. Trust boundaries are the product.

04 / Catalog

What's in the catalog.

URL-level contextual classifications

4 billion URLs graded at the vector level, IAB 3.1 compatible, queryable in any taxonomy.

Brand-safety overlays

Risk scoring, sensitivity gating, and category exclusion lists composable on top of any classification.

Custom taxonomies

Bring your own segment definitions. The vector is the substrate. The taxonomy is a mapping table.

Audience and intent signals

Inferred intent at the URL level, available alongside the contextual classification.

Sustainability and quality grades

Carbon impact, content quality scoring, and ad-density indices for buyers prioritizing premium inventory.

Trust and provenance receipts

ARTF receipts, source authentication, and synthetic-vs-human content grading.

05 / Platform

Four primitives. One marketplace.

/01Index

Classify

Every URL on the open web. Indexed at the vector level. Updated continuously by the ContentGraph pipeline. The substrate underneath everything else.

/02Match

Query

Search the index in any taxonomy. IAB 3.1, brand-safety, custom segments, agent-defined intent. One embedding, many outputs, real-time response.

/03Transact

Exchange

Buy and sell context through AdCP and MCP. Direct integration with the major DSPs and SSPs. No intermediary identity layer, no cookie dependency.

/04Verify

Trust

ARTF receipts on every transaction. Provable provenance, auditable scoring, agent-readable trust signals. Built so any party can verify any claim.

06 / Standards

Built on open protocols.
Not a walled garden.

01 · Protocol

AdCP

Agentic Content Protocol. The transaction layer for context between buyers and sellers. IAB Tech Lab.

02 · Protocol

MCP

Model Context Protocol. The interface every agent uses to call external context. Vectorplatz exposes ContentGraph as an MCP server.

03 · Framework

ARTF

Agent Retrieval Trust Framework. Receipts, provenance, and scoring on every transaction. Auditable by any party.

04 · Taxonomy

IAB 3.1

The taxonomy the buy side already speaks. One vector, every taxonomy, mapped through tables. No re-classification required.

Manifesto

Identity was the unit of advertising for fifteen years.
Context is the unit of the next fifteen.
The collaboration layer has to follow.

Brendan Norman / Founder, Classify

07 / FAQ

Frequently asked.

The agentic data collaboration layer for the open web. It operates as a marketplace where data providers list contextual signals and buyers (agencies, agent platforms, and AI labs) query, transact, and verify through open standards.
By signal type. Contextual classifications, brand-safety overlays, custom taxonomies, audience and intent signals, quality grades, and provenance receipts. Buyers query across the catalog. Providers list within categories. The vector layer underneath is the substrate that makes cross-provider queries coherent.
Identity platforms graph who a person is. Vectorplatz graphs what content means. The two layers are complementary, but in a world where agents replace browsers as the primary buyer of attention, the meaning graph is the layer that scales.
No. Vectorplatz is a marketplace that integrates with the major DSPs, SSPs, and agent platforms. We sit underneath the buying stack, not alongside it.
AdCP for agentic content transactions, MCP for agent retrieval, ARTF for trust receipts, and IAB 3.1 for taxonomy. We participate in IAB Tech Lab working groups on AdCP and CoMP.
Three access tiers. Marketplace access for agencies and brands, plugged into PubMatic and the major DSPs. Standards integration for tech platforms and agents through MCP and AdCP. Licensing partnerships for AI labs and publishers. Pricing is by usage and licensing terms, set per relationship.
No. Buyer queries and licensing data are used solely to fulfill the transaction and produce ARTF receipts. The classification model itself is trained on a separate human-labeled corpus.

The agentic web needs a venue.
Bring data, query the catalog, or join the standard.