Kong takes out the volcano: TyrayScript, MCP-Native Sdk for production building ai Agents with LLM Consulting and real world actions

Kong is open A volcanoan SDK that proves the functionality of the multi-step agent on the coast for all LLMs of Native providers Model Complex Protocol (MCP) Use of tools. The release combines the extensive capabilities of the MCP in it Kong AI Gateway and Incplacing volcano as a developer SDK in the MCP control plane.

• Why Volcano SDK? Because 9 lines of code are faster to write and easier to manage than 100 +.
• WITH Volcano SDK? You may need 100+ Schemas for tool management, context management, provider switching, error management, and HTTP clients.
• With Volcano SDK: 9 lines.

import { agent, llmOpenAI, llmAnthropic, mcp } from "volcano-ai";


// Setup: two LLMs, two MCP servers
const planner = llmOpenAI({ model: "gpt-5-mini", apiKey: process.env.OPENAI_API_KEY! });
const executor = llmAnthropic({ model: "claude-4.5-sonnet", apiKey: process.env.ANTHROPIC_API_KEY! });
const database = mcp("
const slack = mcp("


// One workflow
await agent({ llm: planner })
 .then({
   prompt: "Analyze last week's sales data",
   mcps: [database]  // Auto-discovers and calls the right tools
 })
 .then({
   llm: executor,  // Switch to Claude
   prompt: "Write an executive summary"
 })
 .then({
   prompt: "Post the summary to #executives",
   mcps: [slack]
 })
 .run();

Which volcano provides?

Volcano presents a compact, More API–.then(...).run()-That goes through the middle context between the steps while changing the LLMS with each step (eg. Planning with one model, remove the other). Treats MCP as a primary monitor: Developers Volcano Hand List of MCP servers, and the SDK does Tool acquisition and persuasion by default. Production features include automatic return, Output for each step, Connection Connection For MCP servers, Oauth 2.1 authenticity, too Telemetry traces / distribution metrics to be observed. The project is released under Apache-2.0.

Here it is Important features volcano SDK:

• More API: Create a multi-step workflow in a nutshell .then(...).run() pattern; context flows between steps
• Use a traditional tool: Exit MCP servers; The SDK auto detects and assigns the appropriate tools at each step.
• Multi-Provider LLM support: Organize models (eg planning in one, execution in another) within a single workflow.
• Signing of the intermediate and final effects of the agent's responsive interaction.
• Retrieval and expiration Optimized for each step to be reliable under real-world errors.
• Responsibilities (Before / after the step) to customize the behavior and the Instrument.
• Typo error handling failure is an active failure during the execution of the agent.
• Parallel execution, branches and loops to express a complex control flow.
• Viewing with Opentelemetry With tracking and metrics on channels and tool calls.
• Oauth Support and Connection Pooling Secure, efficient access to MCP servers.

Where found in the MCP design of Kong?

Kong Inc The platform adds multiple MCP management and access layers compatible with Volcano's SDK VORCE:

• AI Gateway SINS MCP Gateway features such as Server Autogeneation from kong-managed apis, Oauth 2.11 For MCP servers, and recognition Tools, workflow, and motivation on connect dashboards. This provides uniform policy and analytics for MCP Analytics.
• This page Konnect Developer Portal can be turned into MCP server so AI coding tools and agents can Discover APIs, Request Access, and use Endpoints Reducing the workflow of the application and making API lists available through MCP.
• Kong's team is also reviewed The creator of MCP and MCP Runner Designing, manufacturing, and deploying MCP servers and integrations.

Key acquisition

• Volcano is open source Typing An SDK that builds multiple AI agents with Use of the First MCP Tool.
• The SDK provides tools for production-Retrieval, Timeout, connection connection, oauthagain Telemetry Tracking/metrics-for MCP Workflows.
• Volcanoes erupt multi-llm Techniques / assassinations and auto-discoveries / invokes MCP servers / toolsto reduce the glue code.
• Kong paired the SDK with platform controls: AI Gateway / Konnect additional MCP Server Autogenation, Oauth 2.1 Intermediate and 2.1, and authentication.

Kong's volcano SDK is an addition to the pragmatic ecosystem of MCP: A first-of-its-kind framework that matches the developer's performance with business controls (OAUth 2.1, Opentelemetry) delivered through AI Gateway and Konchect. Pairing closes a common gap in Agent Stacts discovery—Tool, Author, and Views—without creating new spaces with MCP. This design puts forward the integration of the Protocol-Nation McP with the integration of a bespoke glue, cutting the active flow and the closure of the test spaces as internal dimensions of the agent.

Look Github repo and Technical details. Feel free to take a look at ours GitHub page for tutorials, code and notebooks. Also, feel free to follow us Kind of stubborn and don't forget to join ours 100K + ML Subreddit and sign up Our newsletter. Wait! Do you telegraph? Now you can join us by telegraph.

Michal Sutter is a data scientist with a Master of Science in Data Science from the University of PADOVA. With a strong foundation in statistical analysis, machine learning, and data engineering, Mikhali excels at turning complex data into actionable findings.