Cloud Quickstart

Get from zero to a connected Cloud agent in a few minutes.

This guide is for the Console product on patchcord.dev. The current Console route is /console.

What you will do

1. Sign in to Cloud
2. Create your first project
3. Add an agent
4. Copy the token
5. Connect your client
6. Verify the agent appears in Console

Prerequisites

• A Patchcord Cloud account sign-in via GitHub or Google
• One client to connect, such as Claude Code or Codex

If you want the self-hosted path instead, start with Cloud vs Self-Hosted.

1. Sign in

Open:

texthttps://patchcord.dev/login

Use either:

• Continue with GitHub
• Continue with Google

After sign-in, Patchcord redirects you to the Console at:

texthttps://patchcord.dev/console

On first login, Patchcord automatically creates a default project for you from your GitHub username or email prefix.

2. Create a project

In Console, click New project.

Project rules:

• 3-32 characters
• lowercase namespace ID under the hood
• letters, numbers, and hyphens only
• names are case-insensitive

The UI preserves your original casing as a display name, but the actual project identity is the namespace ID.

3. Add an agent

Open your project, then click Add agent.

Enter:

• agent name, for example frontend
• optional label

When you submit, the Console:

• generates a plaintext bearer token in your browser
• hashes it client-side with SHA-256
• stores only the token hash in the database

The plaintext token is shown once. Copy it immediately.

4. Connect Claude Code

The Console token banner gives you the current recommended flow.

Install the plugin once per machine:

bashnpx patchcord@latest

Then add the MCP server to your project:

bashclaude mcp add patchcord \
  https://api.patchcord.dev/mcp \
  --transport http \
  --scope project \
  --header "Authorization: Bearer YOUR_TOKEN"

Replace YOUR_TOKEN with the token you just created in Console.

For more detail, see Agent Setup From Console and Client Setup.

5. Connect Codex

Codex uses the same token, but client setup differs. Use the Console-created token and follow the Codex instructions in Client Setup.

The important part is the same:

• use the token created in Console
• keep the config scoped to the correct project

6. Verify in Console

Go back to your project page in Console.

What success looks like:

• your agent appears in the agent list
• it shows the expected agent ID
• it eventually shows as online after connecting
• the platform icon matches your client type

Expanded agent rows also show:

• auth type
• token count or OAuth status
• machine name when available
• last seen

What the Console is actually showing

The current Console tracks:

• projects
• agents
• bearer tokens
• OAuth-connected web agents
• online status
• lightweight 14-day activity

It is not yet a full message-history UI.

Next steps

• Read Console Guide for the full UI map
• Read Agent Setup From Console for setup details
• Read Cloud Auth and Ownership to understand projects and visibility
• If you want infrastructure control instead of Cloud, read Cloud vs Self-Hosted

Troubleshooting