Gemmit is an AI-powered development assistant by Tcmartin, designed for the Model Context Protocol (MCP) platform. It functions as an MCP server tool, bridging Google's Gemini CLI with local development environments. Built with Python 3.8+ for backend services and Node.js 18+ for its Electron GUI, Gemmit provides developers with AI capabilities for project management and code generation. It supports cross-platform installers for macOS, Windows, and Linux, distributed via GitHub Releases with auto-updates.

What does Gemmit connect to?

Gemmit connects directly to Google's Gemini CLI, utilizing its AI capabilities for various development tasks. It serves as a Model Context Protocol (MCP) server, acting as an intermediary between the Gemini model and the user's local development environment or other services. This connection allows Gemmit to access AI-powered functionalities without requiring local model downloads. The application's AI-powered chat interface specifically uses WebSockets for real-time streaming conversations, ensuring dynamic interaction with Gemini 2.5 Flash. Gemmit itself runs entirely off Google's Gemini CLI, making the Gemini CLI a primary compatible AI client.

Core features and tools

Gemmit offers several capabilities as an AI-powered development assistant. These include an AI-Powered Chat Interface for real-time streaming conversations with Gemini 2.5 Flash. It provides Project Scaffolding for automatically creating new projects. Users can perform File Operations to browse, create, read, update, and delete files directly within the application. Scoped Intelligence allows the AI to introspect existing codebase and suggest context-aware changes. Additionally, an Extensible CLI Core enables users to utilize the full power of "gemini-cli" and swap in any Gemini-based model.

Who is it for?

Gemmit targets developers seeking an AI-powered development assistant for their coding workflows. It offers "AI vibe coding" to assist with project initialization, code generation, and iterative development. The tool supports the PocketFlow methodology, guiding users through design, implementation, and optimization phases. Its design focuses on integration with Google's Gemini CLI, providing AI assistance without complex setup. This is not the right fit if you require a development tool that operates independently of Google's Gemini CLI.

How to get started

No credentials are required, as Gemmit runs entirely off Google's Gemini CLI without needing API keys or local model downloads. Its AI-powered chat interface uses WebSockets for real-time streaming conversations.

Use cases

• Project Initialization and Scoping: Launch Gemmit, create or point it to a project in "~/Gemmit_Projects/", then let Gemmit analyze the code to suggest next steps.
• AI-Assisted Code Generation and File Management: Interact with the built-in chat UI to prompt Gemini for code, documentation, or tests, then modify, save, or generate new files directly within the application.
• Iterative Development with PocketFlow: Design, implement, and optimize projects following the PocketFlow methodology, supported by Gemmit's AI assistance.

Gemmit by Tcmartin provides an AI-powered development assistant experience, integrating with Google's Gemini CLI. It supports developers with features like project scaffolding, file operations, and context-aware code suggestions. The application offers a free-to-use model without requiring API keys or local model downloads, making it accessible for AI-assisted coding. It is distributed cross-platform with auto-updates. Try using the AI-Powered Chat Interface to generate code or documentation for your current project.