Indrajala (Indra's Net) is a platform for distributed event processing, IoT, and AI integration. It serves as an intelligent event router that connects various data sources, sensors, and processing modules (including LLMs) through a unified protocol.

• Distributed Architecture: Supports Python, C++ (ESP32), Swift (iOS/macOS), and JavaScript/WASM clients.
• Event-Based: Uses the IndraEvent protocol (MQTT-like addressing with UUIDs) for flexible data routing.
• Modular: Easily extensible with new processing modules.
• AI Integration: Built-in support for ML models and LLMs.
• Resilient: Supports asynchronous processing and fault tolerance.

The system consists of:

• IndraServer: The core Python server that routes events between modules.
• IndraLib: Libraries for interacting with the server.
• Client Libraries: Support for multiple languages (Common Lisp, JS, Swift).

• Python 3.13 or higher
• uv (recommended for dependency management) or pip

1. Clone the repository:

   git clone https://codeberg.org/domschl/Indrajala.git
   cd Indrajala

2. Install dependencies:

   pip install -r indraserver/requirements.txt
   # OR if using uv
   uv sync

1. Configuration: The server requires a configuration file. You can start with the default:

   cp config/indrajala.default.toml config/indrajala.toml

   Edit config/indrajala.toml to enable/disable modules and set paths.

2. Run the Server: You can use the provided helper script:

   ./ij.sh

   Or run directly with Python:

   python indraserver/src/indrajala.py

Indrajala uses a specific event structure called IndraEvent. Messages are routed based on domain (like MQTT topics). See doc/indra.md for protocol details.

See LICENSE file.