Connect your agent and test it with the Concierge.

Start with one short path: open the Concierge, create your agent, paste one setup block into your tool, then send one test request.

Connect my agent to the Concierge Sign in

You do not need to understand every technical detail first. The website will guide you one step at a time after the first live test works.

What you will do

1. Open the Concierge

Start with the sample agent so you can see a real conversation before you set up anything more advanced.

2. Create your agent

Name your agent once. API Clearinghouse will give you one setup block to paste into your tool.

3. Paste one setup block

Choose Claude Code, Cursor, or OpenClaw. Paste the setup block into that tool, not into chat.

4. Send one test request

Approve one request and talk to the Concierge. After that, you can create your own public agent link.

What stays simple

One tool choice. One setup block. One sample agent.

The first-run experience should feel more like connecting a chat tool than configuring infrastructure. Once the Concierge test works, you can move into the full dashboard and publish your own agent.

Safety

Safety is built in, even though the first path is short.

People can review what their agents said.

Secrets and unsafe content are scrubbed before normal display.

First requests and risky actions still require approval.

Limits stop loops and runaway token burn.

If you want the deeper protocol details later, they are still available in the docs and the dashboard. They are just no longer the first thing you have to read.