MCP Servers · adhikasp

mcp-client-cli

mcp-client-cli is a Python CLI tool that lets you interact with any LLM (OpenAI, Groq, local models) and connect to MCP-compatible servers for extended capabilities. It provides a terminal alternative to Claude Desktop with support for image analysis, prompt templates, and tool confirmation workflows.

Talk to DEV.co Discuss Implementation

Source: GitHub — github.com/adhikasp/mcp-client-cli

678

GitHub stars

Forks

Python

Primary language

MIT

License (OSI-approved)

Key facts

Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.

Field	Value
Repository	adhikasp/mcp-client-cli
Owner	adhikasp
Primary language	Python
License	MIT — OSI-approved
Stars	678
Forks	83
Open issues	22
Latest release	v1.0.5 (2025-12-02)
Last updated	2025-12-02
Source	https://github.com/adhikasp/mcp-client-cli

What mcp-client-cli is

A Python-based CLI that orchestrates LLM API calls (via provider abstraction) with Model Context Protocol (MCP) server connections over stdio. It implements tool invocation with optional confirmation gates, multimodal input handling, conversation state persistence, and configuration-driven MCP server spawning.

Quickstart

Get the mcp-client-cli source

Clone the repository and explore it locally.

terminalbash

1git clone https://github.com/adhikasp/mcp-client-cli.git2cd mcp-client-cli3# follow the project's README for install & configuration

Need it deployed, integrated, or customized instead? DEV.co ships production installs.

Best use cases

Terminal-first LLM workflows with tool access

Engineers who prefer CLI-based interaction and need to compose MCP tools (fetch, search, APIs) without leaving the terminal. Useful for scripting, CI/CD integration, and headless environments.

Multi-LLM experimentation and comparison

Quickly swap between OpenAI, Groq, local llama.cpp, or custom LLM endpoints via config changes. Ideal for benchmarking responses or using cost-effective providers for specific workloads.

Rapid prototyping of MCP-based agent workflows

Developers building MCP server implementations can test tool integration and conversation flows directly from the CLI before deploying agents. Supports prompt templates for domain-specific tasks.

Implementation considerations

Requires manual ~/.llm/config.json setup with LLM API keys and MCP server command definitions; no wizard or validation tool mentioned.
MCP servers are spawned as child processes via configurable commands (uvx, npx, etc.); caller must ensure dependencies and PATH resolution work correctly.
Tool confirmation workflow requires user interaction; bypass via --no-confirmations flag but may reduce safety in unattended automation.
Conversation state persists locally; no mention of multi-machine sync, backup, or export mechanisms for chat history.
Clipboard feature depends on OS-specific tools (xclip on Linux, pbpaste on macOS, PowerShell on Windows); WSL support adds complexity.

When to avoid it — and what to weigh

Real-time collaborative or multi-user environments — CLI tool is single-user, single-machine focused. No built-in user management, audit logging, or concurrent session handling for team use.
Requirement for guaranteed data privacy or air-gapped deployment — Depends on external LLM providers (OpenAI, Groq) by default; local-only mode via llama.cpp is possible but not the primary design. API keys must be stored in config files.
Need for web UI, mobile, or non-terminal interfaces — CLI-only interface; no web dashboard, API server, or GUI. Organizations requiring a managed SaaS or enterprise UI should look elsewhere.
Production-grade availability or SLAs — Early-stage project (created Nov 2024, v1.0.5 as of Dec 2025). Unknown production readiness, error recovery, or uptime guarantees; 22 open issues suggest ongoing stability work.

License & commercial use

MIT License: permissive, allows commercial use, modification, and redistribution with attribution. No restrictions on use in proprietary or closed-source projects.

MIT License explicitly permits commercial use. However, users must verify compliance with any underlying LLM provider terms (OpenAI, Groq) and MCP server licenses. No warranty or indemnification from the project; commercial users bear liability risk.

DEV.co evaluation signals

Editorial assessment — not user reviews. Directional, with an explicit confidence level.

Signal	Assessment
Maintenance	Active
Documentation	Adequate
License clarity	Clear
Deployment complexity	Moderate
DEV.co fit	Good
Assessment confidence	High

Security considerations

API keys stored in plaintext in ~/.llm/config.json and via environment variables; no encryption or secrets management integration. Tool confirmation feature mitigates unintended tool execution. No audit logging, access control, or rate limiting. MCP servers run as child processes with inherited environment; malicious MCP servers could access LLM API keys and clipboard data. Users must validate MCP server integrity before use.

Alternatives to consider

Claude Desktop

Official MCP client with built-in UI, automatic updates, and Anthropic support. Better for non-technical users and guaranteed compatibility, but less flexible for multi-LLM use.

Cline (VS Code extension)

Editor-integrated MCP client with LLM access. Good for developers already in VS Code; more discoverable than CLI but locked to one editor.

LangChain / LangGraph

Programmatic agent frameworks with broader LLM provider and tool ecosystem support. Steeper learning curve but more flexible for custom automation and integration.

Software development agency

Build on mcp-client-cli with DEV.co software developers

mcp-client-cli enables engineers to quickly prototype MCP-driven agent workflows, swap between LLM providers, and access tool capabilities without leaving the CLI. Evaluate compatibility with your MCP servers and LLM provider terms before production use.

Talk to DEV.co

Related open-source tools

Surfaced by semantic similarity across the DEV.co open-source index.

mcpc mcp2cli Windows-MCP

Related on DEV.co

Explore the category and the services that help you build with it.

MCP Servers AI application development API development Custom software Talk to DEV.co

mcp-client-cli FAQ

Can I use mcp-client-cli with local models?

Yes, via llama.cpp. Configure the llm provider with a custom base_url pointing to your local llama.cpp server instance.

Do I need Claude Desktop or a separate MCP client?

No. mcp-client-cli is a standalone MCP client. It runs MCP servers directly from the CLI without requiring Claude Desktop or other clients.

Are tool calls always confirmed by the user?

Only if specified in the requires_confirmation field per MCP server. Use --no-confirmations flag to skip prompts, but this reduces safety.

How do I add a custom MCP server?

Add an entry to mcpServers in config.json with a command, args, and optional env variables. The server must implement the MCP stdio protocol.

Software developers & web developers for hire

Adopting mcp-client-cli is usually one piece of a larger software development effort. As a software development agency, DEV.co provides software development services and web development expertise — pairing senior software developers and web developers with your team to design, build, and operate mcp servers software in production.

Ready to run LLMs from your terminal?

Talk to DEV.co Schedule a Consultation