Show HN: Henri: a small, hackable agent CLI (github.com)

Henri, a new small and hackable command-line interface (CLI) agent developed in Python, has been launched, inspired by Claude Code. This tool allows users to interact with multiple large language model (LLM) providers, including AWS Bedrock, Google Gemini, Vertex AI, and local options like Ollama. Key features include real-time token streaming, tool integrations for executing commands (like bash and file operations), and a robust permission management system that ensures users can control command execution through prompts. The significance of Henri lies in its clean architecture and flexibility, making it easy for developers to customize and extend. Users can create their own tools and providers by subclassing existing components and defining specific behaviors, promoting a collaborative development environment. Moreover, its permission management allows for secure operations, which is critical for applications involving sensitive data. As a hackable CLI agent designed for experimentation and development, Henri holds potential for advancing the use of AI in various pipeline solutions, appealing to both developers and researchers in the AI/ML community.