TLDR: Kong Inc. has launched Volcano, a new TypeScript-based, open-source SDK designed to simplify and accelerate the creation of Model Context Protocol (MCP)-powered AI agents. Alongside Volcano, Kong announced significant enhancements to its AI Gateway, including new MCP governance capabilities, centralized authentication, improved observability, and an AI Integration Composer and Runner. These innovations aim to address challenges in security, reliability, and performance in agentic AI workflows, enabling developers to build and deploy advanced AI agents more efficiently.
NEW YORK – Kong Inc., a prominent developer in cloud API and AI technologies, has introduced ‘Volcano,’ an innovative, open-source Software Development Kit (SDK) built with TypeScript, specifically tailored for the rapid and secure development of AI agents. The announcement was made at the API Summit 2025, where Kong also revealed substantial upgrades to its AI Gateway, reinforcing its commitment to fostering the ‘agentic AI era.’
Volcano distinguishes itself by being the first SDK to adopt a Model Context Protocol (MCP)-native approach to AI agent creation. According to Mark Palladino, Kong CTO and co-founder, existing frameworks often present complexities and abstractions that hinder the concise and elegant development of AI agents. He highlighted that many older solutions predate the widespread adoption of MCP, a protocol that has emerged as a crucial connectivity layer in the agentic AI landscape, linking AI to various data sources, systems, and services.
During his keynote, Palladino demonstrated Volcano’s capabilities by showcasing an AI agent that, with just 15 lines of code, processed over 900 coffee orders in real-time at the conference. This agent leveraged participant names and locations to auto-generate and place orders at nearby coffee shops, illustrating the SDK’s efficiency and power.
Volcano supports a vast array of models from leading vendors, including Anthropic, OpenAI, Azure, AWS, and GCP. A key feature is its ability to create flexible workflows, allowing different Large Language Models (LLMs) to be utilized for distinct parts of a task. This enables developers to select the most suitable model for specific reasoning or processing steps, optimizing performance. The SDK can also generate subagents that can be natively consumed by other agents, fostering a modular and scalable development environment. Kong plans to expand model support and build an ecosystem of integrations, with the SDK being available on GitHub at https://github.com/Kong/volcano-sdk and more information at https://volcano.dev/.
In conjunction with Volcano, Kong CEO and co-founder Augusto Marietti unveiled new MCP gateway capabilities for Kong’s AI Gateway, which was initially launched last year. The Konnect Developer Portal is now MCP-enabled, offering a unified and secure platform for discovering, building, running, and governing MCP-based tools and applications powered by AI agents.
Key enhancements to Kong’s AI Gateway for MCP governance include:
Consistent MCP Server Autogeneration and Protection: Automated generation of secure MCP servers from Kong-managed APIs, accelerating the deployment of smart AI agents and boosting developer productivity.
Centralized MCP Authentication: A dedicated MCP OAuth 2.1 authentication plugin allows for central security management of all MCP servers, validating tokens and applying consistent security policies automatically.
Stronger MCP Observability and Governance: Enhanced data capture for MCP tools, workflows, and prompts, providing comprehensive analytics in Kong Konnect dashboards for cost optimization, performance monitoring, and improved governance, complementing existing LLM analytics.
AI Integration Composer and Runner: Marking Kong’s entry into AI-based integration, these features enable the discovery, composition, and mapping of intelligent API endpoints as tools for MCP servers, and the generation and execution of code for these integrations, all while ensuring security, reliability, and governance through the Kong AI Gateway.
Also Read:
- Red Hat AI 3 Unveiled: Empowering Enterprise AI with Scalable Production Inference and Agentic Capabilities
- BigID Introduces Model Context Protocol Server for Secure AI-Native Enterprise Data Access
These advancements underscore Kong’s mission to empower companies to become ‘API-first’ and securely accelerate the adoption of AI coding assistants, AI workflows, and agentic AI, addressing critical infrastructure challenges in the evolving AI landscape.
