FLUJO is an open-source platform designed for AI developers and researchers to manage, orchestrate, and interact with multiple AI models and MCP (Model-Context-Protocol) servers locally. It provides a unified, visual interface for building complex AI workflows, managing API keys securely, and integrating with various AI providers and local models, simplifying the development and deployment of sophisticated AI applications.

How It Works

FLUJO leverages the PocketFlowFramework and is built with C++. It offers a visual flow builder for designing workflows, allowing users to connect different models and tools. The platform integrates with MCP servers, enabling users to install, manage, and inspect their tools. It also features a chat interface for interacting with these workflows and supports OpenAI-compatible endpoints for seamless integration with other AI applications.

Quick Start & Requirements

• Installation: Clone the repository (git clone https://github.com/mario-andreschak/FLUJO.git), navigate to the directory, and run npm install or yarn install.
• Running: Start the development server with npm run dev or yarn dev. For a compiled version, use npm run build and npm start. Desktop app development is available via npm run electron-dev.
• Prerequisites: Node.js (v18 or higher), npm or yarn.
• Docker: For easier setup and Docker-based MCP server integration, clone the repo and run docker-compose up -d.
• Documentation: DOCKER.md for detailed Docker information.

Highlighted Details

• Secure storage for environment variables and API keys with encryption.
• Support for multiple AI models, including local integration via Ollama.
• Visual workflow orchestration with node-based design and branching/looping capabilities.
• Direct chat interface for interacting with created flows.
• MCP server integration, including installation from GitHub/local filesystem and Docker support.

Maintenance & Community

• The project is actively maintained by mario-andreschak.
• Community support is available via Discord: https://discord.gg/KPyrjTSSat.

Licensing & Compatibility

• Licensed under the MIT License.
• Permissive license suitable for commercial use and integration with closed-source applications.

Limitations & Caveats

FLUJO is in an early preview state, with features like file attachments and transcription noted as "really bad atm." Extensive logging is enabled by default, potentially exposing API keys in terminal output, requiring user caution.