A New Dawn: Kong Unleashes API Technologies for the AI Era
The main fundamental question of every business today: Are you on the right side of AI?
Many businesses and investors are facing the big question: will they get AI-fied and die slowly or will they be propelled by AI and grow faster?
Kong API tools and the Konnect API platform will bring our customers on to the right side of AI.
Every platform shift from mainframes onward has created demand for more APIs. The PC era only had one with Windows APIs and then Mobile and Cloud had a Cambrian explosion of API endpoints. Now with the advent of AI, we have a whole new game with Agents talking programmatically via endpoints. After all, LLM input and output are APIs.
The future ahead for the API economy is much bigger than anything we’ve seen in the past.
The big takeaway is that we’re in the middle of an interface shift.
We’re moving from an era of User Interface(UI)-based services intended to be consumed by actual people - interacting visually and interfacing with a mouse and a keyboard, over to API-based services that are consumed by AI agents. In fact the Internet as we know it is becoming an enormous neural network to be consumed by machines interacting with each other via APIs.
I admit this is perhaps a bit of a Matrix-influenced view of the future, but the truth is this AI dominated time is coming much faster than what we ever imagined. We’re at the cross road of an Interface shift.
The initial shift is visible among the many LLMs that are taking over our software. It doesn’t matter if they are hosted, large, small, tiny, cloud or custom models. They all talk via APIs.
Thus, the big bet here is that security, observability and controls are going to move from the LLM frameworks and backends into the Traffic Layer. It’s how it happened in microservices and it’s how it’s going to happen in the GenAI era. No one will rewrite a token rate limiting a dozen times in different LLMs or languages.
In the short time since the rise of LLMs, organizations have quickly discovered that GenAI can make a huge impact. This is the fastest adoption of any technology that I've ever seen. Example: 83% of developers report AI investments have already led to new products or services in the last year, and 92% say AI is a priority in the next year, according to the 2024 API Impact Report survey of nearly a thousand developers and IT business leaders.
Don’t believe me? Let's look at this quote from the AI master.
It is for this world that API management is evolving once again. We started the revolution by building the first cloud native API Gateway, then we introduced a Unified API Platform with Konnect with one control plane to manage both North-South and East-West traffic and now we proclaim once more to evolve our API platform into…
The API platform for the AI Era
As mentioned AI traffic is nothing more than API traffic going from and to LLMs. This traffic requires APIs to be built first, those APIs have to run securely and reliably, then APIs and their infrastructure have to be governed, and ultimately the consumer-facing API developer experience has to be delightful.
Two characteristics make it “Intelligent”: it is AI-infused with features such as traffic prediction, smart plugins suggestions and anomaly detection but most of all is an AI enabler for your organization with products like the AI Gateway.
To do AI, you have to do non-AI stuff first. And that’s why an API management that is AI infused, and an AI enabler is essential. From the cataloging of API building blocks for GenAI to complete tasks, to AI Gateway and AI runner as first steps.
This is the API Platform for the AI Era.
Organizations must ensure the performance, security and governance of LLMs usage in applications to be secure and reliable at scale. The solution lies in the relationship between LLMs, AI and APIs. With the critical nature of the intersection of these technologies in mind, I'm thrilled today to unveil many new Kong products and announcements at API Summit that will foster your journey into the AI era.
There is no AI, without APIs.
Onwards!
###
Post Show Round up
Here’s a recap of our latest product releases announced at API Summit 2024 this week, with a range of new capabilities for Kong Konnect to drive all of the build, run, and govern API lifecycle phases.
Build
With Kong Insomnia 10 AI Runner capabilities, developers build GenAI services once and can then instantly access the various LLMs supported by Kong AI Gateway, speeding up their productivity and strengthening code quality. The first AI runner in the industry is a lightweight version of the Kong AI Gateway and provides intelligent multi-LLM infrastructure for AI to every Insomnia developer. Insomnia users can now consume multiple GenAI LLMs by writing their code only once. Out of the box, users can redirect their applications to Azure AI, AWS Bedrock, GCP Vertex, OpenAI, Anthropic, Mistral, and Cohere without any code changes. This accelerates experimentation, benchmarking, and the final implementations of GenAI applications and agents.
In addition to AI Runner, Kong Insomnia 10 introduces Collection Runner and Invite Control to further enhance the developer experience. These features streamline API testing and development workflows, allowing developers to focus on building high-quality and highly performant APIs.
Run
Kong Konnect is a unified API platform that integrates with many runtimes such as API gateways, AI gateways, ingress controllers, and service meshes under one unified experience end-to-end.
We've announced the release of Kong Gateway 3.8 the most powerful API Gateway yet. Some new features coming to the world's most adopted API gateway include Incremental Configuration Sync and OpenTelemetry support — two critical capabilities that optimize the operational efficiency of APIs by reducing latency, conserving resources, and ensuring robust observability across diverse infrastructures. In addition, our upcoming Config Store and Secret Vault further streamline the secure management of API configurations, making it easier for teams to run their services reliably at scale.
Kong AI Gateway, launched in February 2024, enables engineering organizations to build LLM-powered applications faster and with no additional code. With Kong Konnect, it offers a platform as a service to roll out AI development teams faster, while securing, accelerating, cost saving and observing the traffic from and to your LLMs.
New features being announced as part of Kong AI Gateway 3.8 are semantic intelligence capabilities that drive faster development time, load balancing for performance and high availability, lower cost of LLM ownership, and stronger security and governance for GenAI app development.
API and AI gateways can be deployed in a hybrid mode, in the cloud, or fully managed by Kong — fitting the needs of large and small organizations.
We've expanded the Dedicated Cloud Gateway offering to support Microsoft Azure and additional AWS regions to provide a high-performance API infrastructure globally. These updates ensure that organizations can deploy and manage APIs across multiple clouds with enterprise-grade SLAs, reducing the operational burden and allowing teams to focus on innovation rather than infrastructure.
We've also announced Kong Serverless Gateways, a breakthrough in API deployment. By offering instant access to Kong Gateway’s capabilities without the need for data plane management or scaling know-how, this topology accelerates the “get stuff done” for developers and takes out of their way any concerns for running APIs.. It's particularly beneficial for start-ups and small teams launching new ideas, as well as for rapid iteration and deployment.
Govern
We introduced Kong Konnect Service Catalog, which will serve as the de facto system of record for all APIs run in any organization. Service Catalog is important as the ultimate system of record for LLMs and Agents to go find and query the API building blocks to get a task done. Also solves the challenge of “shadow and zombie APIs" — those that exist but are not managed or visible yet. By enabling the discovery, cataloging, and governance of all APIs within an organization, we reduce the risks and increase developer productivity with a centralized view of all API services that are needed to build applications.
Finally, Advanced API Analytics is getting a facelift with new AI Analytics capabilities in Kong Konnect providing precision troubleshooting with real-time, detailed insights into API performance. By offering granular visibility into metrics such as latency and request size — as well as AI LLM-specific reporting metrics — these tools empower teams to maintain and improve API quality and performance across their organization, ensuring APIs meet business requirements consistently such as SLAs.
If you are ready to see what the API platform for AI can do for your organization…try it for yourself. Get started free with Kong Konnect/free.