Best for AI models and management in one place.

What I like most is the federated model: it gives each team its own gateway while still keeping central governance intact. I manage a few microservices across two environments, and before Konnect I was SSHing into different boxes just to push config changes. Now the SaaS control plane lets me handle everything from a single dashboard. I can spin up a runtime, attach my services, and configure rate limiting and JWT auth through the UI, and it’s live within minutes. The developer portal is also a quiet win. I publish my APIs as products with OpenAPI specs, and external teams can self-serve onboarding instead of emailing me for credentials. The plugin system is genuinely modular, too—I’ve stacked key-auth, rate-limiting, and response-transforming plugins on the same route without writing a line of custom code.

Kong Connect gives us a central layer for managing APIs while still fitting into our workflow. We mainly use it with our MCP server, and having the APIs set up so we can manage policies via Kong makes it easier for us to keep the integration controlled, instead of handling everything separately. The AI gateway part is also useful because it can manage authentication and our token usage across all services, rather than adding the same logic to each individual service. Overall, it helps us make better sense of the traffic.