Nasiko provides a centralized platform for managing AI agents at scale, acting as a developer control plane. It addresses the complexities of deploying, routing, and monitoring multiple AI agents by offering an Agent Registry, intelligent query routing, an API Gateway, and comprehensive observability tools. This platform is designed for developers and researchers who need to orchestrate and manage AI agent ecosystems efficiently, enabling seamless integration and intelligent interaction between agents.

How It Works

Nasiko employs a microservices architecture, featuring a FastAPI backend for core logic, Kong API Gateway for routing and load balancing, and an Intelligent Router Service powered by LangChain for sophisticated query analysis and agent selection. Agents are containerized and registered in a central registry, allowing for automatic discovery and deployment onto Kubernetes or local Docker environments. The platform emphasizes observability through OpenTelemetry and LangTrace, providing distributed tracing and performance metrics for agent interactions, which is advantageous for debugging and understanding complex agent workflows.

Quick Start & Requirements

Local development is recommended for initial setup.

• Primary install/run command: Clone the repository, install dependencies using uv sync, and start the platform with make start-nasiko.
• Non-default prerequisites: Python 3.12+, Docker, Docker Compose, Kubernetes (v1.24+ for cloud deployment), and an OpenAI API key (optional, for router service). A native macOS Desktop App is also available for download.
• Links: Quick Start, Documentation, Architecture, Desktop App, CLI Tool, Cloud Deployment.

Highlighted Details

• Intelligent Routing: Automatically routes queries to the most suitable AI agent using LangChain and capability matching.
• Agent Registry & Orchestration: Centralized management, versioning, and deployment of AI agents.
• Observability: Integrated tracing (LangTrace, OpenTelemetry) and metrics for deep insights into agent interactions.
• Developer Tools: Comprehensive CLI tool, SDK, and a web interface for seamless agent management and testing.
• Scalable Deployment: Supports deployment to Kubernetes clusters (AWS, DigitalOcean) and local Docker environments.

Maintenance & Community

Support and bug reporting are managed through GitHub Issues and Discussions. The project is actively developed, with clear contribution guidelines provided.

Licensing & Compatibility

The project is licensed under the MIT License, permitting commercial use and integration with closed-source applications.

Limitations & Caveats

The provided Desktop Application is specific to macOS. While the platform supports Kubernetes and Docker, detailed setup for specific cloud providers or complex network configurations may require manual adjustments beyond the one-command bootstrap.