How to Access Conveyor in Claude, Cursor, OpenAI, and More

The Conveyor MCP server is live

Conveyor now lives inside the AI tools your team already uses. OpenAI Playground, Claude, Cursor, and more. Connect the new Conveyor MCP server once and your knowledge base, trust center, questionnaires, and analytics show up in whichever AI companion you reach for.

If your company is trying to do more with AI right now (whose isn't?), this is the big customer trust unlock. If you're a Conveyor customer, follow the steps below and you'll be running a real Human + Agent customer trust playbook by the end of the day.

But first, a song that's going to be stuck in your head for the rest of the day.

Y, MCP?

Accurate security knowledge is what closes big deals. It's also what holds them up. Most of it lives in the heads of the security team and a knowledge base nobody else can reach. Conveyor fixes that inside our app. The MCP server brings that same power to your favorite LLM. Now your security content shows up in the AI tools the rest of the company is already working in.

Approved users skip the context switch and pull from Conveyor anywhere they're prompting.

What MCP actually is

MCP stands for Model Context Protocol. It's an open standard that lets AI tools talk to outside data and software so they can actually get work done.

Compare that to a traditional API. You need a developer, time, and a roadmap slot. MCP cuts that out. Non-technical users get the same integrations without writing a line of code.

"MCP is a standard plug for AI. Connect it to Conveyor and your assistant suddenly has all of your security knowledge to work with." — Anner, CTO

Once the Conveyor MCP server is connected to your AI tool, you get the same data and integrations that used to require an in-app session or API call. Just in a chat window.

What you can do with it

The Conveyor MCP server connects to our public API, so you get Q&A, trust center management, questionnaire workflows, and analytics through chat or through agents.

The longer you use it, the more it can do. Here's how to think about it, in three levels. We'll use Claude as the example, but it works the same everywhere else.

Level 1: Ask a question, get an accurate answer

Ask Claude a security question. It pulls the answer from your Conveyor Knowledge Library through the MCP. That's a knowledge base your team curated, our AI monitors, and your reviewers signed off on. Not whatever Claude found on the public internet or in a stray Notion doc.

Example: A customer asks, "What specific controls do you have in place to prevent prompt injection or jailbreaking?" Your sales rep types it into Claude, gets the approved Conveyor answer back, and sends it to the customer. No security ticket. No waiting.

Pro tip: Want sales to answer security questions on their own? Build a Claude skill that spots security questions and pulls answers from the MCP, then roll it out to GTM. Self-serve security answers, every deal.

Level 2: Run the customer trust workflow without leaving Claude

Approve pending requests, review knowledge base issues, manage your trust center. All from chat.

Example: A known customer requests access to your trust center. The email hits your inbox. You approve it from Claude. "Approve all pending access requests from @acme.com," set the permissions, done. No tab switching. No logging into a second tool.

Level 3: Run agentic workflows that reference the Conveyor MCP

This is where it gets interesting.

Example: Audit the questions your team is being asked across sales, customers, and your trust center. Then turn that audit into a roadmap to improve your content.

Try this prompt:

I want to analyze what security questions are being asked of our team and our trust center. Review the #security_questions Slack channel and questions asked by sales over the last 60 days. Also review all Conveyor Trust Center interactions over the last 60 days. You can use tools from the Conveyor MCP server for that.

What are the top questions and documents? Which questions from sales aren't answered by the Trust Center? What questions don't have answers? What answers should we work on to make more helpful?

I want you to similarly review the curated Q&As (popularity, expiration, duplicates, etc.) and propose actionable feedback to improve the content and the customer experience.

Summarize it all in a nicely formatted PDF report.

You just audited your trust program with a prompt.

How to connect to the Conveyor MCP server

Head to the MCP start guide.
Generate an API key and point your AI client at https://mcp.conveyor.com/mcp with your token as a Bearer header. Nothing to install. Nothing to self-host.
Start asking questions. Once connected, your assistant automatically discovers what it can do. Try: "What questionnaires are due this week?" or "Which customers downloaded our SOC 2 report in the last 30 days?"

Try this Claude skill with the Conveyor MCP server

The Conveyor MCP server is available now in Labs (beta). Access is open, but we're working on SSO and approval in the OpenAI and Anthropic ecosystems. Once you're connected, drop this prompt into Claude and let it run.

Trust Center Activity Audit

I want to analyze the content on our Conveyor Trust Center. You can use tools from the Conveyor MCP server for that. Let's focus on interactions from the last 60 days.

First, analyze the documents shared there. Which ones are popular? Which ones are unused? Any other insights?

Second, review the curated Q&As (popularity, expiration, duplicates, etc.) and propose actionable feedback to improve the content and the customer experience.