DocumentationIntegrations

Integrations

Claude Desktop

Create a dedicated XEDOC connection for the Code tab.

Connect automatically

#

Install Node.js 20 or later, fully quit the selected client, and run this command. The assistant checks your key, shows the files it will update, and saves a private backup before applying changes.

Windowspowershell
npx.cmd --yes https://xedoc.top/downloads/xedoc-setup-1.0.1.tgz --clients claude-desktop

Open the Code tab

#

Reopen Claude Desktop after setup. The assistant adds and selects a named XEDOC inference configuration while keeping your other saved configurations. This connection is for the Code tab and does not change the standard Claude chat service.

Configure in the app

#
  1. Enable the connection settings

    Open Help → Troubleshooting → Enable Developer Mode if the third-party inference menu is not visible. Restart, then open Developer → Configure Third-Party Inference.

  2. Add a gateway

    Create an XEDOC connection with the values below.

  3. Add Claude models

    Use canonical XEDOC model IDs and choose the matching family tier. Set only one default per family.

  4. Apply and restart

    Apply locally, reopen the app and select a model in the Code tab.

Connection settingstext
Inference provider: Gateway
Gateway base URL: https://api.xedoc.top
Gateway API key: sk-xedoc-YOUR_KEY
Credential kind: Static API key
Gateway auth scheme: Bearer
Suggested modelstext
Model ID: claude-opus-5
Label: Claude Opus 5
Family tier: opus
Family default: On

Model ID: claude-sonnet-5
Label: Claude Sonnet 5
Family tier: sonnet
Family default: On

Managed devices

#

Organization policy can take precedence

An organization-managed inference configuration may override local settings. If your app remains on a managed provider, ask your administrator to configure XEDOC through the supported policy. The assistant does not alter organization policy.

Reference: Claude Desktop configuration.