izan.io

izan.io is an open-source AI assistant platform that aggregates multiple AI providers and MCP tools in a privacy-first browser-centric experience. It exposes a suite of built-in MCP servers and browser automation capabilities that let you record browser actions, extract data via CSS selectors or accessibility tree, and convert those recordings into MCP tools that can be orchestrated by agents. The system supports long-running tasks, scheduled automations, and multi-provider access, enabling you to run complex workflows without exposing your data to the server side. To use it effectively, start the web app locally, add your API keys for the providers you want to use in the Settings panel, and begin by configuring MCP tools or recording new ones with the action recorder extension. You can also chain multiple MCP tools and agents to create multi-step workflows where the output of one tool becomes the input for the next, up to three levels deep.

Prerequisites:

• Node.js 18+ and npm 10+ installed on your machine
• Git installed

Install and run locally:

1. Clone the repository git clone https://github.com/ekingunoncu/izan.io.git
2. Change into the project directory cd izan.io
3. Install dependencies npm install
4. Start the development server npm run dev
5. Open the app in your browser http://localhost:5173

Environment and config:

• See apps/web/.env.example for optional environment variables to customize the web app behavior.
• You can override or add environment variables via a local .env file or by exporting them in your shell before starting the server.

Tips and common issues:

• Privacy-first: API keys are stored in the browser (IndexedDB) and not sent to the server by default. Ensure you provide keys for the providers you plan to use in the Settings panel.
• If the dev server fails to start, verify Node.js and npm versions, then run npm ci to install dependencies cleanly.
• The MCP ecosystem in izan.io includes browser-based MCP servers (mcp-browser-servers) and extension-based MCP servers (mcp-extension-servers). The action recorder can generate MCP tool definitions from recorded browser actions and store them as JSON in IndexedDB or load from S3/CloudFront.
• For production deployments, follow the Deploy instructions in the repo to set up S3/CloudFront or other hosting, and configure domains with ACM certificates as described in the docs.