Show HN: Kdeps – A New CLI Coding Agent That Runs on Open Weight Models (kdeps.com)

Kdeps, a new command-line interface (CLI) coding agent, has been unveiled, allowing developers to run AI workflows entirely offline using open-weight models like Llama 3.2 and Claude-opus-4-8, without the need for API keys or cloud dependencies. Kdeps simplifies the development process by enabling users to create workflows in a single YAML file, deployable across Docker, Kubernetes, or as standalone binaries. This platform-agnostic approach allows developers to switch backends seamlessly without altering their code, which significantly enhances flexibility and reduces lock-in risks associated with traditional AI frameworks. The significance of Kdeps lies in its capacity to streamline the coding process by eliminating boilerplate code and manual dependency management. Developers can compose agents to function like independent microservices, enhancing modularity and reusability. Kdeps not only supports automated error handling but also facilitates orchestration of multiple agents within a unified framework. This tool represents a step forward for the AI/ML community, promoting ease of use and efficiency in creating autonomous AI systems that can be deployed reliably in production environments.