APIs.io has always been about making APIs discoverable by machines. The APIs.json format provides a machine-readable index of what APIs exist and where to find them. That solves the discovery problem for APIs themselves.
But APIs are only half the story. The integrations built on top of APIs — the clients, the wrappers, the middleware that connects system A to system B — are largely invisible. They are not indexed. They are not machine-readable. They live as code in repos that nobody outside the building team knows about.
Spec-Driven Integration changes this. When integrations are declared as structured specifications rather than encoded in application code, they become discoverable artifacts — indexable by APIs.json, searchable by AI agents, and inspectable by governance teams.
Discoverable Integrations
A Naftiko capability spec declares what an integration consumes from upstream APIs and what it exposes to downstream consumers. That spec has a clear identity: a namespace, a description, typed inputs and outputs, consumed sources, and exposed tools. It is a structured artifact that can be indexed exactly the way APIs.json indexes APIs.
When your integration layer is spec-driven, discovery extends beyond “what APIs does our organization have” to “what governed integrations exist, what do they consume, and what do they expose.” That is the discovery surface AI agents actually need.
From API Discovery to Capability Discovery
APIs.json indexes APIs. With SDI, APIs.json can also index capabilities — governed integration units that sit above APIs and serve specific business workflows. The same index format, the same discovery pattern. The unit of discovery evolves from raw API endpoints to governed, shaped, named capabilities that are ready for AI consumption.
This is the natural extension of what APIs.io has always advocated: machine-readable discovery for the full integration stack, not just the API layer.
| Read the SDI methodology | Explore the Naftiko Fleet |