Happy slice of mcp pizzaMCP.pizza

mcp-cli

MCP.Pizza Chef: IBM

This is a stand-alone chat program you run in a terminal window, letting you talk with AI models from OpenAI, Anthropic's Claude, Google's Gemini, or free local models through Ollama, all while giving the AI hands-on access to whatever tools you connect. Attach photos, documents, or code files to a conversation, save and reload past chats, and watch the AI carry out multi-step tasks such as looking things up online or working with your files. Cloud AI providers need a personal account key; the built-in local option needs none.

Coding
Data
Web/Research

Use This MCP client To

Chat with Claude, ChatGPT, or Gemini from my terminal Have an AI search the web and fetch results for me Attach a photo or document and ask questions about it Save a conversation and pick it back up later Run a free local AI model with no account needed Let an AI complete a multi-step task on its own Export a chat as a text file to share

README

MCP CLI - Model Context Protocol Command Line Interface

CI PyPI version

A powerful, feature-rich command-line interface for interacting with Model Context Protocol servers. This client enables seamless communication with LLMs through integration with the CHUK Tool Processor and CHUK-LLM, providing tool usage, conversation management, and multiple operational modes.

Default Configuration: MCP CLI defaults to using Ollama with the gpt-oss reasoning model for local, privacy-focused operation without requiring API keys.

🆕 Recent Updates (v0.16)

AI Virtual Memory (Experimental)

  • --vm flag: Enable OS-style virtual memory for conversation context management, powered by chuk-ai-session-manager
  • --vm-budget: Control token budget for conversation events (system prompt is uncapped on top), forcing earlier eviction and page creation
  • --vm-mode: Choose VM mode — passive (runtime-managed, default), relaxed (VM-aware conversation), or strict (model-driven paging with tools)
  • /memory command: Visualize VM state during conversations — page table, working set utilization, eviction metrics, TLB stats (aliases: /vm, /mem)
  • Multimodal page_fault: Image pages return multi-block content (text + image_url) so multimodal models can re-analyze recalled images
  • /memory page <id> --download: Export page content to local files with modality-aware extensions (.txt, .json, .png)

Execution Plans (Tier 6)

mcp-cli FAQ

Do I need to pay for anything to use this?
No — it can run completely free using a local AI model on your own computer. If you would rather use a cloud AI such as OpenAI's ChatGPT, Anthropic's Claude, or Google's Gemini, you will need a personal account key from that company.
Can I use this to search the web for me?
Yes — once you connect a web search tool, you can ask the AI to look things up online and bring back real results instead of guessing.
Can I use this to work with PDFs, photos, or code files?
Yes — you can attach documents, images, and code snippets to a conversation and ask questions about them.
Which apps does this work with?
It is its own stand-alone chat window that runs in your terminal, rather than a plugin added to Claude Desktop, Cursor, or ChatGPT. It talks directly to AI models from OpenAI, Anthropic, Google, and others, including free local models.
How hard is it to set up?
It takes some comfort with installing command-line programs — you install it with a tool like pip or npx and run it from a terminal window, so it suits someone willing to follow written install steps rather than a one-click install.
Will it remember my past conversations?
Yes — you can save a conversation, close the program, and pick it back up later exactly where you left off.
Can it complete multi-step tasks on its own?
Yes — you can ask it to plan out and carry out a task in several steps, and it will work through each one and show its progress along the way.
Is my information kept private?
If you use the free local model option, everything stays on your own computer. If you use a cloud AI provider, your conversation is sent to that company under their own privacy terms.