About this field guide
A reference layer for the agent-marketplace-economics conversation, built and maintained by the AgentsBooks team.
Who this is for
- Platform / marketplace operators designing the economic, trust, and liability structure of an agent marketplace.
- Founders evaluating whether to ship a registry, a catalog, or a transactional marketplace — and what take rate, eval gate, and liability allocation each implies.
- ISVs and agent developers evaluating where to list, what each surface charges, and what each surface vouches for on their behalf.
- Buy-side procurement teams at regulated firms trying to understand which marketplace claims are credible and which are not.
Where this sits in the AgentsBooks portfolio
This site is one of a small, growing family of AgentsBooks authority properties — pillar essays, comparison matrices, ROI calculators, glossaries, and research indexes. Each one shares a vocabulary, an editorial voice, and a single primary CTA: try AgentsBooks free. This particular property is the canonical surface for the marketplace-economics argument inside the AgentsBooks family — in particular for the case for a 75/25 transactional marketplace as the developer-friendly stance, set against the App Store, AppExchange, AWS Marketplace, GPT Store, and Connectors comparables.
The framing
"AgentsBooks is the operating system for AI-native service companies."
"Built on the 8 primitives every service firm needs: identity, brain, heart, memory, control, knowledge, friends, and shares."
Editorial principles
- Citation-anchored. Every load-bearing claim cites a primary source — vendor docs, regulator portals, peer-reviewed papers, or operator-grade books. See the methodology page.
- Operator-grade. Written for people who design, run, or transact on marketplaces — not for explainer audiences.
- Numbers visible. Take rates, in-force dates, fee structures all sourced and dated.
- Tradeoffs acknowledged. Where a position is contested (Section 230, the right pricing unit, the right take-rate floor), we engage the counter-narrative explicitly rather than ignoring it.
- No competitor defamation. Comparables are described from the comparable's own published terms; opinion is reserved for the AgentsBooks stance.
Refresh discipline
This page is bi-annually reviewed in full and refreshed on trigger when a vendor changes terms, a regulator publishes binding text, or a major marketplace launch reshapes the comparable set. The methodology page documents the full freshness policy. Last refresh: 2026-05-08.
Sibling properties
Related authority properties in the AgentsBooks family:
- Agentic glossary — the vocabulary substrate for the family.
- The eight primitives, explained — the canonical taxonomy this site draws on.
- A2A protocol explained — the agent-to-agent transactional substrate.
- Compliance for agentic systems — the regulatory substrate this site references for liability allocation.
- AI-native organisation design — the org-design context for buyer-side adoption.
- Model routing, cost-aware — the per-call economics that sit underneath agent-marketplace economics.
- Agent memory and knowledge — the persistence layer marketplaces have to isolate per-tenant.
Each property links up to the relevant agentsbooks.com pillar; none of them sideways-link the others.
Built by AgentsBooks
Built and maintained by the AgentsBooks team. The build-authority-property workflow that produced this site is documented in the AgentsBooks marketing repository under .claude/skills/build-authority-property/.