Swagger Meeting You Where You Work

Some approaches to API governance interrupt developers mid-flow, forcing them to context-switch into a separate tool and manually verify their API definition before they can ship. That approach has never really worked. Not because developers don’t care about quality, they do, but because the best time to fix an API is the moment you’re already thinking about it.

That’s what has always guided how Swagger grows. Not “come to us.” But “we’ll be there.”

It started with a definition

When the OpenAPI Specification took off, the Swagger open-source tooling and SwaggerHub became the place you designed and documented APIs. It was a destination: you opened Swagger, you designed or created your API, you published your docs. That worked because API design was a deliberate, separate activity. You sat down to do it.

Then the world changed. APIs stopped being something teams designed in isolation. They became the connective tissue of entire organizations, shipped fast, updated often, consumed by services, agents, and third-party developers who depended on them not changing unexpectedly. Governance couldn’t live in a separate tab anymore.

The Swagger tooling meets developers where they are working, such as in their IDEs, into the review step, not just the planning step.

API work has moved into the AI assistant

Developers don’t context-switch to a governance tool anymore. They run a check from the same interface where they just wrote the code. Product managers don’t open a separate tab to review an API definition. They ask their AI assistant.

This is why Swagger is now available as a Claude connector, on GitHub, and as a Kiro Power. These are the places where API work happens today.

What it looks like in practice

Here’s a real example. A developer is working in Claude and needs to ship an updated train booking API. Instead of opening Swagger, running a governance scan, reading the violations, fixing them manually, and saving a new version, they type one sentence:

“Find the train-booking-api in the swaggerclaudeconnect organization, validate it against governance rules, fix any issues, and save it as a new version.”

Claude finds the API, runs it through the organization’s Spectral ruleset, surfaces the violations, applies the fixes, and saves the corrected API definition as a new version. The developer stays in the flow. The API ships clean.

That’s not a demo. That’s what governance looks like when it lives where you work.

The same goes for documentation. Instead of maintaining docs separately from the API definition, the perennial “we’ll fix it later” that never happens, a PM can ask Claude to generate portal documentation and publish it directly. The API definition and the docs stay in sync because they’re part of the same conversation.

Where Swagger is now

• Claude connector available in the Anthropic Connectors Directory. Connect your SmartBear account once and your APIs, governance rules, and documentation portals are accessible through natural conversation inside Claude. No installation, no tokens to manage.
• GitHub the open-source SmartBear/smartbear-mcp server, published as @smartbear/mcp. Works in VS Code and GitHub Copilot. The same governance and design capabilities, wherever your team already works.
• Kiro Power the Swagger Contract Testing Power for AWS’s Kiro IDE. Connects to a PactFlow workspace for can-i-deploy checks, provider verification, and consumer impact analysis. Contract testing as a natural part of the development loop, not a gate at the end of it. Learn more in our blog post.

Why this is bigger than it looks

This isn’t just a productivity tool. It represents something more significant: Swagger is now part of the AI assistant ecosystem. By building a Remote MCP Server and listing it on Anthropic’s Connectors Directory, SmartBear opened the door to a new distribution channel. Any Claude user, in Claude.ai, in Claude Code, in Claude Desktop, can now find Swagger as a Claude Connector and connect with a single click. No installation, no configuration, no friction.

And because the protocol is open and shared, one Swagger MCP Server can surface in Claude, Copilot, Kiro, and whatever agent ships next. This is how AI-native workflows are built. Not by replacing tools, but by bringing them into the conversation.

What’s next

The pattern here is consistent: wherever API work moves, Swagger moves with it.

That means more AI coding platforms beyond Claude, Copilot, and Kiro. It means richer Kiro Powers, agents and skills that handle more of the API lifecycle without manual steps. And it means functional testing MCP tools, so an agent can validate how an API behaves, not just whether it conforms to an API definition.

APIs are increasingly written by agents and consumed by agents. The tools that govern them need to speak the same language. That’s what we’re building toward: Swagger is available for execution wherever the work is, with the same rules applied consistently every time.

The Swagger MCP Server is available now. Install from github.com/SmartBear/smartbear-mcp, add the connector in Claude, or try the Kiro Power in the Kiro IDE.