Connect Your Agent

Connect Claude Desktop, ChatGPT, Microsoft Copilot, or any MCP-compatible client to Skilder via MCP.

Skilder acts as an MCP server. Your AI agent connects to it the same way it connects to any MCP server — via STDIO or Streamable HTTP. Once connected, the agent automatically discovers the capabilities (skills and tools) available through its assigned hats.

The Connect dialog in Skilder generates the right configuration for your client. Click Connect from the top navigation bar or from any hat's detail page.

Step 1: Open the Connect Dialog

Click the Connect button in the top navigation bar (or from a hat card). The dialog shows:

"Connect to this MCP Server" — with client options below
Connection Settings (expandable) — select an API key and optionally preload specific hats or skills

If you don't have an API key yet, you can create one directly from the dialog.

Step 2: Choose Your Client

Claude Desktop (STDIO)

The fastest way. Skilder generates a .mcpb bundle file that configures Claude Desktop automatically.

Select Claude in the Connect dialog.
Select or create an API key.
Optionally preload specific hats or skills.
Click Download .mcpb and open the file — Claude Desktop configures itself.

That's it. Open Claude Desktop and start chatting. Your agent discovers all available capabilities automatically.

ChatGPT (Streamable HTTP)

Select OpenAI ChatGPT in the Connect dialog.
Copy the Streamable HTTP Endpoint URL shown (includes your key and any preload parameters).
In ChatGPT settings, add a new MCP connection and paste the endpoint URL.

Microsoft Copilot (Streamable HTTP)

Select Microsoft Copilot in the Connect dialog.
Copy the Streamable HTTP Endpoint URL.
In Copilot Studio, add a new MCP connection with the endpoint URL.

Other MCP Clients

Select Other Clients to see both connection methods:

STDIO — Configure your client to run the Skilder runtime:

{
  "command": "npx",
  "args": ["-y", "@skilder-ai/runtime"],
  "env": {
    "USER_KEY": "your_key_here"
  }
}

Streamable HTTP — Point your client at the endpoint URL shown in the dialog:

https://your-skilder-instance/mcp?key=your_key_here

Both methods support optional hatIds and skillIds parameters to preload specific capabilities.

Step 3: Test the Connection

Send a message to your agent. It should discover and use the skills available through its hats. Try:

What capabilities do you have available?

The agent should list the skills and their tools.

If it doesn't work, check:

The API key is valid and not revoked
The hat has skills assigned with tools linked
MCP servers are in Connected status (check Workspace Settings > MCP Servers)

How the MCP Connection Works

Your Agent ←→ MCP (STDIO or Streamable HTTP) ←→ Skilder ←→ MCP Servers (tools)

Your agent connects to Skilder as an MCP server.
Skilder authenticates via the API key and resolves which hats, skills, and tools are available.
The agent discovers capabilities through MCP's standard tool listing.
When the agent calls a tool, Skilder routes the request to the right MCP server, executes it, and returns the result.

Your agent never talks directly to the underlying MCP servers — Skilder handles routing, permissions, and execution.

Using Studio Instead

If you want to test without an external client, use the built-in Studio. Click Studio in the top navigation — it connects directly to your workspace's capabilities with no extra setup. You just need an AI provider configured.

Next Steps

Quickstart — Full walkthrough of building your first skill and connecting.
API Keys — Understand key types, scoping, and best practices.
Studio — Test skills interactively without an external client.
Troubleshooting — Common connection issues and fixes.

On this page