This project provides a terminal-based AI chat interface accessible via SSH, allowing users to interact with various AI models directly from their command line. It targets developers and power users seeking a lightweight, accessible AI interaction method without needing a graphical interface.

How It Works

The system utilizes a Node.js backend with the SSH2 library to handle SSH connections. The UI is built with React and Ink, rendering a terminal-native interface. Data persistence is managed by PostgreSQL and Redis, though fallback options exist for local development. This approach enables a familiar SSH experience for AI interaction, leveraging existing terminal tools and workflows.

Quick Start & Requirements

• Install/Run: Deploy via Docker using docker compose up -d. Access via ssh username@host -p 22.
• Prerequisites: Docker, Node.js (for local development).
• Configuration: Requires .env file setup for server name, public access, rate limiting, model configuration (API keys, base URLs), and database connections.
• Docs: GitHub Repository

Highlighted Details

• Supports multiple AI models including Gemini and DeepSeek.
• Configurable rate limiting and IP blacklisting/whitelisting.
• Optional support for chain-of-thought reasoning models.
• Includes Umami analytics integration.

Maintenance & Community

The project is actively maintained by ccbikai. Sponsorships are acknowledged, notably from V.PS for server infrastructure. Links to Telegram and X (formerly Twitter) are provided for community engagement.

Licensing & Compatibility

The repository does not explicitly state a license in the provided README. Compatibility for commercial use or closed-source linking is not specified.

Limitations & Caveats

Windows and Linux terminal compatibility are pending user feedback. The project relies on external AI model APIs, requiring valid API keys and potentially incurring costs. The .env configuration is extensive and critical for proper operation.