y-cli 🚀已精选

What is y-cli?

y-cli is a command-line interface chat application that brings AI conversations to your terminal. It runs on any system with Python and uv, and is designed for developers and power users who want to interact with multiple AI models and MCP tools from a terminal environment.

How to use y-cli?

Install y-cli with uvx y-cli or uv tool install y-cli. After installation, run y-cli init to configure your OpenRouter API key and default settings. Start a chat with y-cli chat. Use commands like y-cli bot add, y-cli mcp add, and y-cli prompt add to configure bots, MCP servers, and prompts.

Key features of y-cli

• Flexible storage: local JSONL files or Cloudflare KV/R2 for cloud backup
• Interactive chat interface with tool execution visualization
• Supports multiple bot configurations (any base_url/api_key/model)
• Works with OpenAI and Dify chat completion streaming APIs
• Supports reasoning models like Deepseek-r1 and OpenAI o3-mini
• MCP integration: multiple server configs (stdio/SSE) and persistent daemon

Use cases of y-cli

• Chat with various AI models (OpenAI, Anthropic, OpenRouter, etc.) from the terminal
• Integrate external tools via MCP servers (e.g., web search, file access)
• Switch between different bot configurations without leaving the CLI
• Store and share chat conversations using local JSONL or cloud storage
• Run a persistent MCP daemon for long‑running tool integrations

FAQ from y-cli

What are the prerequisites for y-cli?

You need uv (Python package manager) and an OpenRouter API key. Follow the official uv installation guide and create an API key at openrouter.ai/settings/keys.

How do I install y-cli?

Run uvx y-cli to use without installation, or uv tool install y-cli to install globally. Then run y-cli init and y-cli chat to start.

Does y-cli support MCP (Model Context Protocol)?

Yes, y-cli has fully integrated MCP client support. You can add multiple MCP server configurations (stdio or SSE protocols), manage them with y-cli mcp commands, and run a persistent daemon with y-cli daemon.

What AI providers and model types does y-cli support?

y-cli supports any base_url/api_key/model combination that uses OpenAI‑compatible chat completion streaming or Dify chat‑messages streaming. It also supports reasoning models like Deepseek‑r1 and OpenAI o3‑mini.

What storage options are available for chat history?

Chat history can be stored locally as JSONL files (easy to sync) or in the cloud using Cloudflare KV and R2 for backup and sharing.