Blog/Tray is now in the Claude Connectors Directory. Here's why it's different.

Tray is now in the Claude Connectors Directory. Here's why it's different.

Tray connects Claude to your whole stack, with the governance and control enterprises actually need.

Simon Cole
Simon Cole
6 min read
Published:
Updated:
Tray is now in the Claude Connectors Directory. Here's why it's different.

Most connectors in the Claude Connectors Directory solve a specific problem. Slack gives Claude access to Slack. GitHub gives Claude access to GitHub. That's useful, and Claude can reason its way across separate tools to get work done.

But enterprise processes can’t be left to chance; they need more control. They span dozens of systems, teams, tools, and approval chains. And when those processes are critical, deal desk approvals, compliance checks, order management, employee onboarding, "Claude will figure it out" isn't good enough. You need determinism. You need control. You need a full audit trail. Not to mention, some apps just aren’t MCP-ready.

That's what Tray brings to the Claude Connectors Directory. Today, Tray joins Slack, Asana, GitHub, Figma, Snowflake, and dozens of others in the official, curated collection of verified MCP connections for Claude at claude.com/connectors. But what Tray offers is a fundamentally different class of tool.

One connection. Your whole stack. Governed.

When Claude connects to the Tray Agent Gateway, it doesn't just gain access to a single service. It gains access to 700+ connectors and the ability to run sophisticated, multi-step business processes with the kind of control enterprises require.

Here's what that looks like in practice. A single MCP tool call can trigger an entire order management process: inventory checks, payment validation, fulfillment routing, and exception handling. That logic, that sequencing, those rules, all defined by you, are baked directly into the tool. Claude executes the process. You define exactly what that process does, how it behaves, and what happens at every step.

Not approximately. Every time.

That matters for two reasons beyond the obvious. First, when more intelligence is built into the tool itself, Claude doesn't have to reason through every step, which means fewer back-and-forth calls, faster outcomes, and meaningfully lower token usage. Second, every tool published through Tray carries its controls built in from the start, not bolted on afterward. IT governs what Claude can and can't do before a single workflow goes live.

Activate in minutes. Run with confidence.

Everything above, the cross-stack reach, the determinism, the governance, comes with activation that's a few clicks.

Find Tray in the Claude Connectors Directory, hit connect, authenticate via OAuth, and you're in. Any MCP service your team has published in Tray and you have access to is immediately available in Claude. No IT project or configuration. Not even a single ticket. Just activate and go.

Governance and ease of activation. You need both.

Here's a pattern that plays out constantly in enterprise AI rollouts. A team finds a tool that's easy to activate, starts using it without guardrails, and IT spends the next quarter cleaning up shadow AI, ungoverned data access, and processes nobody can audit.

Ease without governance is just another name for risk.

So IT locks things down. Approval processes, procurement cycles, and configuration requirements that take weeks. Adoption dies. Nobody uses the tools because getting to them is too hard.

So now you've got governance without ease. That's just a more controlled path to the same shadow AI problem you started with.

You have to have both. That's exactly what Tray in the Claude Connectors Directory delivers. Teams can connect in minutes and start activating sophisticated workflows directly in Claude. IT publishes tools with guardrails already built in, controlling what Claude can access, what it can do, and what gets logged, before anyone runs a single process.

What you can actually do with it

Run your most critical processes exactly as designed

Some processes are too important to leave to interpretation. Deal desk approvals, compliance checks, finance sign-offs. These need to follow your policy every time, with the right routing, the right validation, and a full audit trail.

With Tray, you build that logic directly into the MCP tool. Claude executes it in a single request. The workflow handles the complexity underneath: branching approval paths, policy enforcement, notifications, escalations. The outcome is predictable because the process is defined, not improvised on the fly.

"Run our deal desk approval workflow, apply our standard policy, and route for the right level of sign-off."

Reach every system in your stack, including the ones nobody else can touch

Not every business system has a native Claude integration, and many never will. Legacy ERPs, custom internal tools, specialist SaaS platforms built for a specific industry or workflow. With Tray's 700+ connectors, you can bridge Claude to any of them. Build the integration in Tray, publish it as an MCP tool, and it's immediately available in Claude. Your users get a seamless experience against systems that would otherwise be completely out of reach.

And if a connector doesn't exist yet, Tray CDK is the fastest way to build one. You can even use Claude Code to do it. The same AI you're building tools for can help you extend the platform it runs on.

"Look up Acme Corp in our legacy ERP and give me a summary of their outstanding invoices and recent order history."

Connect in minutes. Control everything.

The processes that matter most to your business deserve more than a single-service MCP connection. Find Tray in the Claude Connectors Directory at claude.com/connectors, connect in minutes, and start building MCP tools that run across your entire stack, on your terms.

Loading...