DocumentationGet API key
All integrations

Claude Desktop

Chat apps

The Claude app for macOS and Windows (chat, Cowork, Code) on UltraGPT's Claude models, as a third-party inference gateway.

Website
Base URL
https://api.ultragpt.pro
API format
Anthropic Messages
Models
Ids, prices & context
Fill the guide with your values

Stays in this browser tab — never sent or stored by us. Sign in to create one

Setup

  1. 1In Claude Desktop open Help → Troubleshooting → Enable Developer Mode (on Windows the ☰ menu on the sign-in screen).
  2. 2Open Developer → Configure Third-Party Inference… and, under Connection, set Inference provider to Gateway.
  3. 3Gateway base URL https://api.ultragpt.pro (without /v1), your UltraGPT key as the API key, auth scheme Bearer.
  4. 4Click Apply Changes — the app relaunches. Claude models are discovered from /v1/models automatically.
Inference providerGateway
Gateway base URLhttps://api.ultragpt.pro
Gateway API key<your UltraGPT key>
Auth schemeBearer
{
  "inferenceProvider": "gateway",
  "inferenceGatewayBaseUrl": "https://api.ultragpt.pro",
  "inferenceGatewayApiKey": "<your UltraGPT key>",
  "inferenceGatewayAuthScheme": "bearer"
}

Good to know

  • The model picker only lists models whose ids are recognizably Claude (for example claude-opus-4.8).
  • Developer → Configure Third-Party Inference → Export produces the same settings as a .mobileconfig / .reg / JSON file for rolling out to a team.
  • To go back to a claude.ai account, switch Inference provider back in the same window, or turn off Developer Mode.
Tips
  • Any model id from Models works; suffixes like :online (web search) and :nitro work too.
  • Give each tool its own key with a spend limit, so usage shows separately in Activity.
  • Something off? Error codes · Support