AI marketplace
Also known as: agent marketplace, AI app store, model marketplace, AI plugin store
Marketplaces in AI follow the same platform logic as app stores: aggregate supply (tools, models, agents) with demand (users, developers, enterprises looking for solutions), take a transaction fee or platform cut, and benefit as the ecosystem scales. OpenAI's custom GPT store was an early large-scale example; MCP server registries and agent capability directories are emerging versions.
For builders, AI marketplaces represent both a distribution opportunity and a dependency risk. Being listed in a high-traffic marketplace can drive discovery and adoption without a full sales team. But marketplace rules, curation decisions, and fee structures are controlled by the marketplace owner, not you. The platform can delist, change terms, or launch a competing product.
The most interesting emerging form is agent tool marketplaces, where individual capabilities, meaning things an AI agent can call or do, are listed as composable services. An agent planning a task can browse and invoke tools the same way a developer browses npm packages (a public repository of reusable code modules). This creates a monetization path for specialized tool builders who don't need to build a full product, just a high-quality, well-documented capability.