Nevron - Lightweight AI Agent Framework

Build flexible autonomous AI agents effortlessly. Set up in 2 minutes, run in 5, deploy in 10.

Quick Start

Prerequisites

• Python 3.11+
• Node.js 18+ (for dashboard)
• Poetry (pip install poetry)

Installation

# Clone the repository
git clone https://github.com/axioma-ai-labs/nevron.git
cd nevron

# Install Python dependencies
make deps

# Install dashboard dependencies
make dashboard-deps

Running Nevron

# Start everything (Agent + API + Dashboard)
make dev-full

This starts:

• Dashboard: http://localhost:5173
• API: http://localhost:8000
• Agent: Running in background

Configuration

All configuration is done through the Dashboard UI - no .env file needed!

1. Open the dashboard at http://localhost:5173
2. Go to Settings page
3. Configure your LLM provider and API key
4. Set up agent personality and goals
5. Enable integrations as needed

Configuration is saved to nevron_config.json and loaded automatically.

The dashboard API requires NEVRON_API_KEY. make dev / make dev-full bind to 127.0.0.1 and,
when NEVRON_API_KEY is not set, run with auth disabled for local development. See
Dashboard docs.

Architecture

Nevron uses a decoupled architecture where the API and Agent run as independent processes:

┌─────────────────┐     ┌──────────────────┐     ┌─────────────────┐
│   Dashboard     │────▶│   FastAPI        │────▶│  Shared State   │
│   (Svelte)      │     │   (Port 8000)    │     │  ./nevron_state │
│   Port 5173     │◀────│   + WebSocket    │◀────│                 │
└─────────────────┘     └──────────────────┘     └────────┬────────┘
                                                          │
                              nevron_config.json ─────────┤
                                                          │
                                                          ▼
                                                 ┌─────────────────┐
                                                 │  Agent Runner   │
                                                 │  (Independent)  │
                                                 └─────────────────┘

Benefits:

• Configure everything from the UI
• Restart API without stopping the agent
• Multiple dashboards can connect simultaneously
• Agent continues running if dashboard disconnects

Dashboard Features

Overview Page

• Agent status (running/paused/stopped)
• Start, pause, stop controls
• Real-time statistics
• Live event feed via WebSocket

Activity Page

• Agent cycle history (Plan → Execute → Learn → Remember)
• Detailed cycle inspection
• Performance metrics

Settings Page

• LLM Configuration - Provider, model, API key
• Agent Settings - Personality, goal, behavior
• Actions - Enable/disable available actions
• Integrations - Configure Twitter, Discord, Telegram, etc.
• MCP Servers - Add custom tool servers

Other Run Options

Available Commands

# Development
make deps            # Install Python dependencies
make dashboard-deps  # Install dashboard dependencies
make format          # Format code
make lint            # Run linting
make test            # Run tests

# Running
make run             # Run agent (legacy mode)
make run-agent       # Run agent process (decoupled)
make api             # Run API server
make dashboard       # Run dashboard
make dev             # Run API + dashboard
make dev-full        # Run agent + API + dashboard

# Docker
make docker-up       # Start all services
make docker-down     # Stop all services
make docker-logs     # View logs

Supported LLM Providers

Integrations

Nevron supports various external services (configure in Settings):

• Social Media: Twitter, Discord, Telegram, Slack, WhatsApp
• Development: GitHub
• Storage: Google Drive
• Media: Spotify, YouTube
• Web3: Lens Protocol
• Search: Tavily, Perplexity

Documentation

For detailed documentation, visit: https://axioma-ai-labs.github.io/nevron/

License

This project is licensed under the Nevron Public License (NPL). See the LICENSE file for details.

Made with love by Neurobro