Skip to main content
This guide walks you through setting up your first project, connecting a server, creating a navigator, and running your first query — all in under 5 minutes.

​
Prerequisites

  • A Caylex account (sign up at app.caylex.ai)
  • Python 3.10+ installed
  • An MCP server to connect (or use one from the Caylex Catalog)

​
Setup

1

Create a project

In the Caylex dashboard, navigate to Projects and click Create Project. Give it a name (e.g., “My First Project”) and an optional description.A project is an isolation boundary for your deployment — it contains its own servers, navigators, and user credentials.
2

Add a server from the catalog

Go to the Servers page and click Add Server. Browse the Caylex Catalog and select a server to add (e.g., GitHub, Slack, or Notion).Once added, go to your project and connect the server by adding it as a Server Instance. This links the server to your project.
3

Create a navigator

Navigate to the Navigators page and click Create Navigator. Provide a name and description (e.g., “My Assistant”).Then connect the navigator to your project — this creates a Navigator Instance. On the navigator instance page, click Create API Key and copy the generated key. You will need this to authenticate with the Caylex Navigator.
Store your API key securely. It is only shown once at creation time.
4

Set up user authentication

If your server requires authentication (most do), create an Auth Link for your project. Go to the project’s Auth Links section and create a new link. Visit the link yourself to authenticate with the server using your credentials.See Auth Links for a detailed guide.
5

Connect your agent

Install your preferred agent framework and connect to Caylex. Here is an example using the OpenAI Agents SDK:
pip install openai-agents
main.py
import asyncio
import os

from agents import Agent, Runner
from agents.mcp import MCPServerStreamableHttp

async def main() -> None:
    api_key = os.environ["CAYLEX_API_KEY"]
    user_email = os.environ["CAYLEX_USER_EMAIL"]

    async with MCPServerStreamableHttp(
        name="Caylex",
        params={
            "url": "https://proxy.caylex.ai/mcp",
            "headers": {
                "x-api-key": api_key,
                "x-user-email": user_email,
            },
        },
    ) as server:
        agent = Agent(
            name="Assistant",
            model="gpt-5.2",
            instructions="You are a helpful assistant that can access external systems using Caylex.",
            mcp_servers=[server],
        )

        result = await Runner.run(agent, "What tools do you have access to?")
        print(result.final_output)

asyncio.run(main())
Set the environment variables and run:
export CAYLEX_API_KEY="your-api-key-here"
export CAYLEX_USER_EMAIL="you@example.com"
python main.py
Your agent should discover all available tools from the servers connected to your project and list them.

​
Next steps

Core Concepts

Understand the Project, Server, and Navigator model in depth.

More Frameworks

See integration examples for LangChain, Claude SDK, and Claude Desktop.

Configure Permissions

Restrict which tools your navigator can access.

Analytics

Monitor tool usage, errors, and performance.