Application programming interface management company Kong Inc. is expanding support for autonomous artificial intelligence agents with the latest release of Insomnia, its open-source API development platform.

Kong says the latest release, Insomnia 12, is smarter, faster and more accessible for developers building APIs and Model Context Protocol servers, having added native MCP clients, AI mock servers and AI-native commit suggestions.

MCP servers have become essential for AI agents to operate. They serve as integration layers that allow agents to connect with different data sources and third-party tools, so they can access information in real time and perform tasks effectively. The MCP protocol standardizes how agents interact with external resources.

At present, AI agents have a lot of promise, but their adoption is still limited. One reason for this is that developers struggle with building MCP servers, due to a lack of tools that can quickly and reliably validate and test what they’ve built. This is what the Insomnia platform is trying to address, and it does this by extending its familiar and easy to deploy “test-iterate-debug” workflow to AI-native development.

Kong has long established itself as a leader in enterprise-grade API development and management, and it believes there’s an opportunity to extend its leadership to AI. After all, APIs play a key role in accessing AI models and enabling those models to access third-party services and data. And it believes that a lot of what applies to API development can be extended to MCP servers.

With Insomnia 12, Kong says developers will now be able to test and debug new MCP servers using the test-iterate-debug workflow that’s already widely used for API testing. They can simply connect it to their servers, manually invoke a tool, prompt or resource with custom parameters, and then inspect the protocol-level and authentication messages and responses to ensure they’re working properly.

The new release also makes it simple for developers to create a fully functional mock AI server by describing what they need in natural language and providing a specific URL, JavaScript Object Notation sample or OpenAPI specification. With this capability, Kong says it can automate hours of manual set up work in a matter of seconds, so developers can spin up mock AI servers for faster testing of their APIs and MCP servers.

Other new capabilities in Insomnia 12 include AI-powered commits, which allows users to automatically create clear, descriptive commit messages that describe the changes they’ve made to a codebase, to help maintain a clean Git history. In addition, users can now choose between local or cloud-hosted large language models, so they have more control over where their AI data and code resides.

Paul Nashawaty, a senior analyst with SiliconANGLE Media’s sister organization theCUBE Research, said Insomnia 12 helps extend Kong’s API development platform into the AI-native era, giving developers a way to test, debug and deploy agentic AI infrastructure much faster while simultaneously maintaining governance, collaboration and code hygiene.

“Industry context from my research shows that over 90% of enterprises see AI agents as inevitable and developer-centric, AI-augmented platforms are among the fastest-growing segments,” he said. “Insomnia 12 addresses critical workflow gaps, mock server setup, protocol-level testing, and version control automation, positioning it as a strategic tool for enterprises moving from AI experimentation to production.”

Kong co-founder and Chief Technology Officer Marco Palladino dropped by SiliconANGLE Media’s mobile livestreaming studio theCUBE last month during the Kong API Summit, where he discussed the critical role of MCP servers for AI agents:

Palladino believes MCP has become the de facto standard for AI agents to communicate, and so it’s essential for developers to have a fast, simple and reliable way to test, debug and govern it. “We’re extending the same developer-friendly workflows we’ve built for APIs to the AI-native world,” he said. “We’re empowering organizations to accelerate agentic AI infrastructure development with built-in confidence, collaboration, and scalability.”

Image: SiliconANGLE/Dreamina AI