Claude Desktop Transforms into Universal Agent Shell: Third-Party Inference Launched, Community Successfully Integrates GPT and Grok

According to monitoring by Beating, Anthropic has launched the Third-Party Inference feature in Claude Desktop. Inference requests from the Cowork and Code tabs can be forwarded to Amazon Bedrock, Google Vertex AI, Microsoft Foundry, or a custom Gateway. The official documentation positions this feature as an enterprise deployment solution for compliance needs: it is suitable for regulated enterprises that cannot send data to Anthropic, international organizations with data residency requirements, and public sector entities operating under FedRAMP or ITAR frameworks. Configuration is delivered through MDM tools, and conversation data is only sent to user-specified inference endpoints and local machines, bypassing Anthropic servers. This feature does not require an Anthropic account; users can access the configuration panel via the top menu Help > Troubleshooting > Enable Developer Mode. Some users have utilized the Gateway interface to connect to third-party gateways like OpenRouter, setting the Gateway base URL to openrouter.ai/api, allowing them to call non-Claude models such as GPT-5.4 and Grok 4.20 in the Cowork interface. This represents a community extension of the Gateway interface and is not an official feature designed by Anthropic for non-Claude models. Testing shows that the 3P mode only provides the Cowork and Code tabs, while the Chat tab, Skills, and Plugins store are unavailable in this mode, and the network exit must be manually configured with a whitelist.