Pull to refresh
Logo
Kong expands Volcano agent toolkit into a full AI-native platform

Kong expands Volcano agent toolkit into a full AI-native platform

New Capabilities

The API connectivity company's bet that AI agents need governed connections now covers model calls, tool use, and agent-to-agent traffic.

Yesterday: Kong expands Volcano into a full AI-native platform

Overview

Updated 33 minutes ago

Kong's open-source AI-agent toolkit became a full platform on September 30, 2026. The San Francisco company announced Volcano as an AI-native way to build, deploy, and run AI agents and web applications. The same day it added native agent-to-agent (A2A) traffic handling, so Kong now manages all three AI traffic patterns: model calls, MCP tool use, and agents talking to each other.

The move extends a bet Kong made in October 2025: that AI agents will reach company data through Model Context Protocol (MCP), an open standard for how models call external tools, not direct, ungoverned API calls. Kong wants to be the governed layer between agents and the enterprise, and a new Portal MCP Server lets agents discover Kong-managed APIs. Whether developers adopt Volcano or stick with bigger frameworks from OpenAI and LangChain remains the open question.

Why it matters

If Kong's bet holds, corporate data reaches AI agents only through governed, observable connections, not direct model calls.

Questions about this story

Free account needed to ask — your question is kept and asked for you right after sign-up. Answers are public.

No questions yet — be the first to ask.

Key Indicators

9
Lines of code for a three-step agent workflow
Kong's SDK demo builds a query, summary, and Slack-post agent in nine lines of workflow code.
thousands
AI models supported out of the box
Volcano works with models from Anthropic, OpenAI, Azure, AWS, and GCP, among others.
3
AI traffic patterns governed natively
Kong AI Gateway handles model calls, MCP tool use, and agent-to-agent (A2A) traffic in one product.

Voices

Curated perspectives — historical figures and your fellow readers.

Ever wondered what historical figures would say about today's headlines?

Sign up to generate historical perspectives on this story.

People Involved

Organizations Involved

Timeline

October 2025 September 2026

4 events Latest: Yesterday
Tap a bar to jump to that date
  1. Kong expands Volcano into a full AI-native platform

    Latest Product Launch

    Kong announces Volcano as an AI-native platform for building, deploying, and operating AI agents and modern web applications, extending the open-source SDK into broader infrastructure.

  2. Kong AI Gateway adds native agent-to-agent traffic handling

    Product Launch

    With AI Gateway 3.14, the Agent Gateway pattern natively manages all three AI traffic flows: model calls, MCP tool use, and agent-to-agent (A2A) traffic. The release also added token exchange, scope-based tool filtering, and support for Databricks, DeepSeek, and vLLM.

  3. Kong introduces Portal MCP Server for agent API discovery

    Product Launch

    The Portal MCP Server turns any Konnect Developer Portal into an MCP server, letting AI agents and coding assistants find Kong-managed APIs and request access in natural language instead of scraping documentation.

  4. Kong unveils Volcano SDK and MCP gateway support

    Product Launch

    At API Summit 2025, Kong open-sources the Volcano TypeScript SDK for building AI agents with MCP tools and adds MCP gateway capabilities to its AI Gateway.

Scenarios

1

Volcano Becomes a Go-To Agent Framework

Possible Resolves by Q3 2027

Discussed by: The New Stack and developer community coverage of the SDK launch

Kong keeps Volcano open-source and lightweight, and its MCP-native design plus multi-provider chaining (OpenAI, Anthropic, Mistral in one workflow) attracts developers who want simple agents without heavy framework dependencies. Growth appears in GitHub adoption and named production deployments rather than marketing pushes.

2

Volcano Stays Niche Behind Incumbent Frameworks

Possible Resolves by Q3 2027

Discussed by: Analysts comparing Volcano with LangChain and OpenAI's Agents SDK

Volcano draws early adopters but fails to dent the lead of larger frameworks. Kong shifts emphasis to its MCP gateway and Konnect governance features, treating the SDK as a reference implementation rather than a growth product.

3

Kong's Gateway, Not the SDK, Wins the Enterprise

Possible Resolves by Q1 2027

Discussed by: Kong's own positioning and enterprise security concerns about direct model API calls

Enterprises adopt Kong's MCP gateway for authentication, rate limits, and observability while writing agents in other frameworks. The platform becomes governed infrastructure rather than a developer framework — the higher-margin outcome for Kong, which has argued that direct API calls by models are insecure.

4

A2A Traffic Governance Becomes Kong's Enterprise Wedge

Uncertain Resolves by Q1 2027

Discussed by: Kong's Agent Gateway release positioning

As agents begin calling other agents directly, enterprises route that traffic through Kong for the same authentication, metering, and observability they apply to model calls. Kong's native A2A handling becomes the reason large firms buy the gateway, even if they build their agents in other frameworks.

Historical Context

2 moments from history that rhyme with this story — and how they unfolded.

2007-2018

The REST API wave and its gateways (2007-2018)

As companies moved data online, REST APIs became the standard way applications talked to each other. A wave of gateway and integration firms — Kong, Apigee, MuleSoft — built the pipes, security, and policy layers around those APIs. Consolidation followed: Google bought Apigee for $625 million in 2016, Salesforce bought MuleSoft for $6.5 billion in 2018.

Then

Enterprise software standardized on governed REST APIs with authentication, rate limiting, and analytics.

Now

Gateway and integration vendors became core infrastructure of the digital economy, worth billions in acquisitions.

Why this matters now

Kong is replaying the playbook for AI: MCP servers as the new endpoints, and gateways that secure and meter agent traffic.

2010-2020

Kong's first act: the API gateway market (2010s)

Kong built its business on open-source API gateway software, routing and securing enterprise traffic as microservices and mobile apps multiplied. The company rode the API economy from startup to a platform now spanning connectivity and management.

Then

Enterprises routed growing volumes of API traffic through Kong's gateway and Konnect platform.

Now

Kong positioned itself as a connectivity provider and began extending into AI traffic management as models emerged.

Why this matters now

Volcano extends that same connectivity business into AI agents — the company's bet on a second, larger wave.

Sources

(11)