gaoze1998/zero-code-cli

Overview

zero-code-cli is a terminal-based AI coding agent built in Rust, designed to streamline the development lifecycle directly from the command line. It addresses the friction of switching between IDEs and browser-based AI interfaces by providing a persistent, context-aware assistant that operates within your local terminal environment. By leveraging the DeepSeek API, it allows developers to execute complex coding tasks without leaving their workflow.

Key Features

• Dual-Mode Workflow: Supports both "Plan" and "Build" modes, allowing users to outline architectural changes before committing to code generation.
• ReAct Tool Loop: Utilizes a reasoning-and-acting loop to iteratively solve problems, enabling the agent to perform multi-step tasks effectively.
• Streaming TUI: Features a responsive terminal user interface that provides real-time streaming of AI responses, keeping the developer informed throughout the generation process.
• Session Persistence: Maintains per-project state, ensuring that the AI retains context across different terminal sessions and coding tasks.
• Memory Safety: Built in Rust, the tool prioritizes performance and stability, providing a robust foundation for automated coding assistance.

Use Cases

This tool is ideal for developers who prefer a keyboard-centric workflow and want to integrate AI assistance into their existing CLI-heavy development process. It is particularly useful for rapid prototyping, generating boilerplate code, debugging complex logic, or managing refactoring tasks where project context is essential. Because it persists sessions per project, it is well-suited for long-running tasks or maintaining consistency across large codebases.

Getting Started

To get started, ensure you have Rust and Cargo installed on your system. You can install the tool by cloning the GitHub repository and building it from source using cargo build --release. Once installed, you will need to configure your DeepSeek API credentials as an environment variable to enable the agent's core functionality. Refer to the repository's README for specific configuration instructions and command-line usage patterns.