Package Exports

This package does not declare an exports field, so the exports above have been automatically detected and optimized by JSPM instead. If any package subpath is missing, it is recommended to post an issue to the original package (@indiccoder/mentis-cli) to support the "exports" field. If that is not possible, create a JSPM override to customize the exports field for this package.

Readme

Mentis-CLI 🧠

An Agentic, Multi-Model CLI Coding Assistant.

Mentis is a powerful terminal-based AI coding assistant that lives in your command line. It supports multiple LLM providers (Gemini, Ollama, OpenAI), agentic file operations, MCP (Model Context Protocol), and more.

✨ Features

🤖 Multi-Model Support: Switch seamlessly between Gemini, Ollama (Local), OpenAI, and GLM-4.6 (Z.AI Coding).
🛠️ Agentic Capabilities: Mentis can read, write, list files, and search your codebase to understand context.
🌐 Web Intelligence: Hybrid search for documentation and error fixing.
🗺️ Smart Context: Automatically maps your repository structure.
🧩 Interactive & Robust CLI:
- History: Navigate previous commands with Up/Down arrows (persistent across sessions).
- Cancellation: Press Esc to instantly stop generation.
- Markdown: Beautiful syntax highlighting in terminal output.
💰 Cost Awareness: Real-time token usage and cost tracking.
🧠 Persistent Memory: Auto-saves sessions (/resume) and checkpoints (/checkpoint).
🔌 MCP Support: Full Model Context Protocol client.
🔍 Codebase Search: Built-in grep tool.
🐚 Shell Integration: Run shell commands and commit changes directly.

🚀 Installation

Using NPM (Recommended)

You can install Mentis globally directly from the NPM registry:

npm install -g @indiccoder/mentis-cli

From Source (Dev)

git clone https://github.com/CoderVLSI/Mentis-CLI.git
cd Mentis-CLI
npm install
npm run build
npm link

⚙️ Configuration

Start Mentis by typing:

mentis

Setting up Models

Type /model inside the CLI to launch the interactive configuration wizard:

Select Provider: Gemini, Ollama, or OpenAI.
Select Model: e.g., gemini-2.5-flash, llama3.
Enter Credentials: API Key (for cloud) or Base URL (for local).

Credentials are stored securely in ~/.mentisrc.

📖 Usage

Modes

/plan: Switch to high-level planning mode (Architecture, requirements).
/build: Switch to code generation mode (Implementation).

Commands

Command	Description
`/help`	Show available commands
`/model`	Configure AI provider and model (Interactive)
`/resume`	Resume the last session
`/checkpoint`	Manage saved sessions (`save`, `load`, `list`)
`/search <query>`	Search for code in the current directory
`/mcp connect <cmd>`	Connect an MCP Server (e.g., `npx -y @modelcontextprotocol/server-time`)
`/run <cmd>`	Execute a shell command
`/commit [msg]`	Stage and commit changes to Git
`/add <file>`	Add a specific file to context
`/drop <file>`	Remove a file from context
`/clear`	Clear chat history
`/exit`	Save and exit

Example Workflow

Mentis > /plan
[PLAN] > Analyze this project structure and suggest improvements.

Mentis > /build
[BUILD] > Implement the FolderManager class in src/utils.ts.

🔒 Safety

Mentis asks for explicit approval before writing or modifying any files, keeping you in control.

🤝 Contributing

Pull requests are welcome!

📄 License

ISC