Add Agentline to your agent

Claude (claude.ai and Claude Desktop)

Add to Claude

1. Click Add to Claude. Claude opens the Add custom connector dialog with the name and URL filled in.
2. Check the URL is https://agentline.nz/mcp and confirm.
3. Click Connect, sign in to Agentline and approve access.
4. Team and Enterprise: an Owner adds the connector under Organisation settings first.

Auth: OAuth (sign in to Agentline). Mechanism checked against claude.com.

ChatGPT

1. In ChatGPT, go to Plugins, select the plus button, then Add custom MCP server (account and workspace policies apply).
2. Name it Agentline. Under Connection, enter https://agentline.nz/mcp as the MCP server URL and choose OAuth for authentication.
3. Review the risk warning, select I understand and want to continue, then Create as a plugin.
4. Sign in to Agentline (or create a free account) and approve access. You're returned to ChatGPT; type @ and pick Agentline in a chat.

Auth: OAuth (sign in to Agentline). Mechanism checked against developers.openai.com.

Claude Code

claude mcp add --transport http agentline https://agentline.nz/mcp

1. Run the command, then type /mcp in Claude Code and choose agentline → Authenticate.

Auth: OAuth, or add --header "Authorization: Bearer al_live_…" for an API key. Mechanism checked against docs.claude.com.

Cursor

Add to Cursor

Config: ~/.cursor/mcp.json

{
  "mcpServers": {
    "agentline": {
      "url": "https://agentline.nz/mcp"
    }
  }
}

1. Click Add to Cursor and confirm the install prompt.
2. Cursor opens the Agentline sign-in the first time a tool needs it.

Auth: OAuth. Mechanism checked against cursor.com.

VS Code (GitHub Copilot)

Add to VS CodeInstall via vscode.dev

code --add-mcp '{"name":"agentline","type":"http","url":"https://agentline.nz/mcp"}'

Config: .vscode/mcp.json (or MCP: Open User Configuration)

{
  "servers": {
    "agentline": {
      "type": "http",
      "url": "https://agentline.nz/mcp"
    }
  }
}

1. Click Add to VS Code. If your browser doesn't open VS Code, use Install via vscode.dev or run the command.
2. Approve the install, then start the server from the MCP view and sign in when prompted.

Auth: OAuth. Mechanism checked against code.visualstudio.com.

Windsurf

Config: Cascade → MCP → Open MCP config file (mcp_config.json)

{
  "mcpServers": {
    "agentline": {
      "serverUrl": "https://agentline.nz/mcp"
    }
  }
}

1. Paste the config into mcp_config.json, save, and press Refresh in the MCP panel.
2. Complete the browser sign-in on first use.

Auth: OAuth. Mechanism checked against docs.windsurf.com.

Cline

Config: Cline → MCP Servers → Configure → Configure MCP Servers (cline_mcp_settings.json)

{
  "mcpServers": {
    "agentline": {
      "type": "streamableHttp",
      "url": "https://agentline.nz/mcp",
      "disabled": false,
      "autoApprove": []
    }
  }
}

1. Or use the Remote Servers tab: name agentline, URL https://agentline.nz/mcp, transport Streamable HTTP, then Add Server.
2. Keep "type": "streamableHttp" — without it Cline falls back to legacy SSE.

Auth: OAuth. Mechanism checked against docs.cline.bot.

Goose

Add to Goose

Config: ~/.config/goose/config.yaml

extensions:
  agentline:
    name: Agentline
    type: streamable_http
    uri: https://agentline.nz/mcp
    enabled: true
    timeout: 300

1. Click Add to Goose (Goose Desktop) and confirm, or add the YAML to your config.

Auth: OAuth. Mechanism checked against github.com.

Zed

Config: settings.json (zed: open settings file), or Settings → AI → MCP Servers → Add Remote Server

{
  "context_servers": {
    "agentline": {
      "url": "https://agentline.nz/mcp"
    }
  }
}

1. Paste the config (note: context_servers, not mcpServers).
2. Click Authenticate next to agentline when Zed asks.

Auth: OAuth (leave out the Authorization header so Zed runs OAuth). Mechanism checked against zed.dev.

Codex CLI

codex mcp add agentline --url https://agentline.nz/mcp && codex mcp login agentline

Config: ~/.codex/config.toml

[mcp_servers.agentline]
url = "https://agentline.nz/mcp"

1. Run the command; codex mcp login opens the Agentline sign-in.
2. For an API key instead: --bearer-token-env-var AGENTLINE_API_KEY.

Auth: OAuth, or an API key via bearer_token_env_var. Mechanism checked against learn.chatgpt.com.

Gemini CLI

gemini mcp add --transport http agentline https://agentline.nz/mcp

Config: ~/.gemini/settings.json

{
  "mcpServers": {
    "agentline": {
      "httpUrl": "https://agentline.nz/mcp"
    }
  }
}

1. Run the command, then /mcp auth agentline inside Gemini CLI to sign in.

Auth: OAuth (auto-discovered), or --header "Authorization: Bearer al_live_…". Mechanism checked against github.com.

n8n

1. Add an AI Agent node, then attach an MCP Client Tool node to its Tools input.
2. Endpoint: https://agentline.nz/mcp. Server Transport: HTTP Streamable.
3. Authentication: Bearer Auth with a credential holding your Agentline API key (al_live_…), or MCP OAuth2 with dynamic client registration on.
4. Tools to Include: All, or pick the ones the workflow needs.

Auth: API key (Bearer) or MCP OAuth2. Mechanism checked against docs.n8n.io.

Any MCP client (JSON)

{
  "mcpServers": {
    "agentline": {
      "type": "http",
      "url": "https://agentline.nz/mcp",
      "headers": {
        "Authorization": "Bearer ${AGENTLINE_API_KEY}"
      }
    }
  }
}

1. Streamable HTTP at https://agentline.nz/mcp. OAuth 2.1 discovery starts from the 401's WWW-Authenticate header.
2. Or skip OAuth with an API key from the dashboard: Authorization: Bearer al_live_….
3. No account yet? Agents can call signup on the keyless server at https://agentline.nz/public/mcp with their human's email.

Auth: OAuth 2.1 or API key. Mechanism checked against modelcontextprotocol.io.