Dell Technologies and Hugging Face have announced a guide for deploying Anthropic's Claude Code and the open-source OpenCode terminal agents on-premises using Dell Enterprise Hub. The guide enables users to run these models directly on their Dell PowerEdge platforms, ensuring data sovereignty and full control over model versions. The setup involves deploying models on the Dell Enterprise Hub, installing the respective CLI tools, and pointing them to the local endpoint, eliminating the need for translation layers or proxies. This approach allows for a fully air-gapped environment where inference occurs on local GPUs behind the firewall, ensuring no data leaves the data center. The guide provides step-by-step instructions for both agents, with detailed commands for deployment and configuration. Users can choose between different model variants based on their hardware and workload requirements, with options for single workstations, team servers, and high-capability systems. The setup also includes guidance on managing credentials and project conventions to streamline workflows. The integration of open-weight frontier models with a best-in-class agent CLI offers benefits such as predictable latency, capacity-based costs, and model auditability. The guide emphasizes the importance of selecting the right model and hardware combination to optimize performance and resource usage. The deployment process is designed to be straightforward, with each command presented in its own code block for easy copying and execution. The guide also highlights the use of vLLM and SGLang for serving models, exposing the standard OpenAI REST API for compatibility. The inclusion of 'Goodput' scenarios with per-GPU service-level objectives (SLOs) helps users size their deployments effectively. The guide concludes with practical steps for running the agents and managing configuration files, ensuring a seamless on-premises deployment experience. The integration of these models with Dell hardware and Hugging Face's platform provides a flexible and secure environment for coding and agentic software engineering tasks. The guide is intended for developers and organizations looking to leverage the power of frontier models while maintaining control over their data and infrastructure. The setup is designed to be scalable, with options for different deployment sizes and model configurations to meet varying needs. The guide also includes tips for managing project conventions and credentials, ensuring a secure and efficient workflow. The integration of these models with Dell hardware and Hugging Face's platform provides a flexible and secure environment for coding and agentic software engineering tasks. The guide is intended for developers and organizations looking to leverage the power of frontier models while maintaining control over their data and infrastructure. The setup is designed to be scalable, with options for different deployment sizes and model configurations to meet varying needs. The guide also includes tips for managing project conventions and credentials, ensuring a secure and efficient workflow.
The guide outlines a five-step process for deploying and configuring both agents, with each step involving specific commands for model deployment, CLI installation, endpoint configuration, credential management, and project setup. The deployment of models on Dell Enterprise Hub involves logging into the platform, selecting a coding model, and running a container command to serve the model on the local server. The installation of the CLI tools for both agents is straightforward, with commands provided for downloading and executing the installation scripts. The configuration of the agents involves setting up the base URL to the local endpoint and specifying the model name, with different configuration files for each agent. Credential management is handled separately for OpenCode, with a dedicated file for storing API keys. The project setup includes creating guidance files for each agent, which define project conventions and environment settings. The guide emphasizes the importance of selecting the appropriate model and hardware configuration to optimize performance and resource usage, with recommendations for different deployment scenarios. The integration of these models with Dell hardware and Hugging Face's platform provides a flexible and secure environment for coding and agentic software engineering tasks. The guide is intended for developers and organizations looking to leverage the power of frontier models while maintaining control over their data and infrastructure. The setup is designed to be scalable, with options for different deployment sizes and model configurations to meet varying needs. The guide also includes tips for managing project conventions and credentials, ensuring a secure and efficient workflow.
The source text describes a practical guide for running Anthropic's Claude Code and the open-source OpenCode terminal agents on-premises using Dell Enterprise Hub. It outlines a five-step process for deploying and configuring both agents, with each step involving specific commands for model deployment, CLI installation, endpoint configuration, credential management, and project setup. The guide emphasizes the importance of selecting the appropriate model and hardware configuration to optimize performance and resource usage, with recommendations for different deployment scenarios. The integration of these models with Dell hardware and Hugging Face's platform provides a flexible and secure environment for coding and agentic software engineering tasks. The guide is intended for developers and organizations looking to leverage the power of frontier models while maintaining control over their data and infrastructure. The setup is designed to be scalable, with options for different deployment sizes and model configurations to meet varying needs. The guide also includes tips for managing project conventions and credentials, ensuring a secure and efficient workflow.
Source: huggingface
