AutoCLI is a command-line tool that fetches information from websites and integrates with local CLI tools through a single unified interface.
The tool solves the problem of accessing real-time data from diverse web sources without writing custom scrapers for each site. It works by providing built-in adapters for over 55 websites including Twitter/X, Reddit, YouTube, HackerNews, Bilibili, Zhihu, and Xiaohongshu, alongside support for controlling Electron desktop applications and integrating existing CLI tools like gh, docker, and kubectl. The implementation reuses browser sessions and employs AI-native discovery capabilities to minimize overhead.
AutoCLI suits developers who need to programmatically gather data from multiple web sources or who want to expose their local CLI tools through a unified command interface. It is particularly valuable for AI agents that require real-time web access, since the tool can be configured in agent configuration files to allow automatic discovery of available commands. The project is a complete rewrite in Rust of an earlier TypeScript implementation, offering substantial performance improvements: it delivers up to 12 times faster execution on certain commands, uses 10 times less memory, and ships as a single 4.7MB binary with zero runtime dependencies, compared to the original Node.js version.
Development activity shows consistent feature expansion with recent additions including a Chrome extension selector tool for visual CSS selector generation, AI-powered adapter discovery that automatically expands scraped fields, and seamless synchronization between local adapters and a cloud marketplace. The project maintains near-feature parity with its predecessor while achieving significant performance gains, as evidenced by test pass rates remaining comparable despite the architectural shift.
