Levante is a free MCP server tool developed by Levante Hub, built with TypeScript. Distributed as a binary desktop application, it serves as a personal, secure, and local AI client. Levante connects to various AI providers, including cloud services and local models, while also integrating with MCP servers to access tools, prompts, and resources. It supports macOS, Windows, and Linux, offering a unified interface for diverse AI interactions. This client focuses on privacy and control for individuals and teams.

What does Levante connect to?

Levante functions as an MCP client, connecting to MCP servers to discover and utilize tools, prompts, and resources. It integrates with numerous AI providers such as OpenRouter, Vercel AI Gateway, and direct cloud services like OpenAI, Anthropic, Google, Groq, xAI, and Hugging Face. Additionally, Levante supports local AI models including Ollama, LM Studio, and custom HTTP endpoints, offering broad compatibility for AI interactions.

Core features and tools

Levante operates as an MCP client, enabling users to interact with tools, prompts, and resources provided by connected MCP servers. Within its chat interface, users can discover, configure, and manage these server-supplied tools. Key capabilities include multi-provider AI interaction, multimodal chat supporting text, images, and audio, and direct integration with MCP servers through an MCP Store & MCP-UI flows for accessing server-based assets.

Who is it for?

Levante is designed for a broad audience, including non-technical users and those who use AI seriously at work, prioritizing privacy and control. It helps teams connect models, tools, and workflows in one unified environment. The program supports individuals and groups seeking to manage diverse AI interactions and server-based resources effectively. This is not the right fit if you require explicit documentation on specific transport protocols for MCP server connections.

How to get started

Access to AI providers is granted through API keys. Initial setup involves connecting Levante to a primary AI provider, such as OpenRouter, by pasting the respective API key into the application.

Use cases

• Multi-provider AI interaction, allowing users to chat with many AI models from various providers through a unified interface.
• Multimodal chat, enabling conversations involving text, images, and optional audio, with automatic routing to appropriate vision or audio models.
• MCP server integration and tool utilization, making MCP tools, resources, and prompts directly accessible within the chat environment.

Levante provides a secure, local desktop client for managing diverse AI interactions and integrating with MCP servers. It supports multiple AI providers and multimodal chat, making server-based tools and resources accessible. The application prioritizes user privacy and control for both individuals and teams. Try connecting Levante to your preferred local AI model, such as Ollama, to explore its capabilities.