Aria: Your Personal AI Assistant

Aria is a powerful, self-hosted AI assistant designed for reasoning, task completion, and seamless interaction. It provides a feature-rich chat interface with multi-session support, real-time streaming responses, and advanced AI capabilities.

✨ Features

• Multi-Session Chat: Manage multiple conversations with persistent history.
• Real-Time Streaming: Get instant responses from the AI as they are generated.
• Markdown & Mermaid Support: Render rich text and diagrams directly in the chat.
• AI-Powered Tools:
  • Web Search: Access real-time information from the web.
  • Reasoning Engine: Perform complex reasoning and analysis.
  • YouTube Analysis: Extract insights from YouTube videos.
  • Weather & Finance Data: Get up-to-date information on weather and financial markets.
• Prompt Improvement: Automatically enhance your prompts for better AI responses.
• Dockerized Deployment: Easy to set up and run with Docker.

🏗️ Architecture

Aria is built with a modern, modular architecture:

• Backend: A robust FastAPI server that handles all business logic, including:
  • Session and message management
  • Integration with Ollama for AI model access
  • A suite of AI tools for enhanced capabilities
• Frontend: A responsive web UI built with HTML, CSS, and JavaScript, providing a seamless user experience.
• AI Core: Powered by Ollama, allowing you to use a variety of open-source language models.
• Services:
  • Redis: For caching and session management.
  • SearXNG: For private and secure web search.
  • Byparr: For additional backend functionality.

🚀 Getting Started

Prerequisites

• Docker
• Docker Compose
• An accessible Ollama instance

Installation

The recommended way to install Aria is by using the official Docker image.

1. Create a docker-compose.yml file:

   networks:
     mitty: null
   
   volumes:
     redis: null
   
   services:
     aria:
       image: 'ghcr.io/malvavisc0/aria:latest'
       container_name: aria
       hostname: aria
       restart: always
       env_file:
         - ./.env
       ports:
         - '8000:8000'
       environment:
         TZ: '${TZ}'
         SEARXNG_URL: '${SEARXNG_URL:-http://searxng:8080}'
         BYPARR_URL: '${BYPARR_URL:-http://byparr:8191/v1}'
         BYPARR_TIMEOUT: '${BYPARR_TIMEOUT:-120}'
         AGNO_TELEMETRY: '${AGNO_TELEMETRY:-false}'
         OLLAMA_URL: '${OLLAMA_URL}'
         OLLAMA_MODEL_ID: '${OLLAMA_MODEL_ID}'
         OLLAMA_MODEL_TEMPERATURE: '${OLLAMA_MODEL_TEMPERATURE:-0.65}'
         OLLAMA_MODEL_CONTEXT_LENGTH: '${OLLAMA_MODEL_CONTEXT_LENGTH:-1280}'
         DEBUG_MODE: '${DEBUG_MODE:-false}'
       volumes:
         - './data:/opt/storage'
       networks:
         - mitty
     redis:
       image: 'redis/redis-stack-server:latest'
       container_name: redis
       hostname: redis
       restart: always
       runtime: '${RUNTIME:-runc}'
       expose:
         - '${REDIS_PORT:-6379}'
       environment:
         TZ: '${TZ}'
       volumes:
         - 'redis:/data'
       networks:
         - mitty
     searxng:
       image: 'docker.io/searxng/searxng:latest'
       container_name: searxng
       hostname: searxng
       restart: always
       volumes:
         - './searxng:/etc/searxng'
       environment:
         TZ: '${TZ}'
         SEARXNG_BASE_URL: 'http://${SEARXNG_HOSTNAME:-searxng}:${SEARXNG_PORT:-8080}/'
         SEARXNG_SECRET: '${SEARXNG_SECRET}'
         UWSGI_WORKERS: '${SEARXNG_UWSGI_WORKERS:-4}'
         UWSGI_THREADS: '${SEARXNG_UWSGI_THREADS:-4}'
         SEARXNG_REDIS_URL: 'redis://redis:6379/10'
       expose:
         - '${SEARXNG_PORT:-8080}'
       cap_add:
         - CHOWN
         - SETGID
         - SETUID
       networks:
         - mitty
     byparr:
       image: 'ghcr.io/thephaseless/byparr:latest'
       container_name: byparr
       hostname: byparr
       restart: always
       runtime: '${RUNTIME:-runc}'
       environment:
         TZ: '${TZ}'
       expose:
         - '${BYPARR_PORT:-8191}'
       networks:
         - mitty
     dozzle:
       image: 'amir20/dozzle:latest'
       container_name: dozzle
       hostname: dozzle
       ports:
        - 8080:8080
       volumes:
         - '/var/run/docker.sock:/var/run/docker.sock:ro'
       environment:
         - DOZZLE_NO_ANALYTICS=1
         - DOZZLE_HOSTNAME=aria
         - DOZZLE_LEVEL=debug
       networks:
         - mitty

2. Configure your environment:

   • Create a .env file and add your configuration. At a minimum, you need to set OLLAMA_URL.

     TZ=Europe/Berlin
     OLLAMA_URL=http://your-ollama-instance:11434
     OLLAMA_MODEL_ID=cogito:14b
     SEARXNG_SECRET=your-super-secret-key

3. Run the application:

   docker-compose up -d

4. Access Aria:

   • Open your web browser and navigate to http://localhost:8000.

⚙️ Configuration

You can configure Aria using the following environment variables in your .env file:

Variable	Description	Default
TZ	Timezone for the application	Europe/Berlin
OLLAMA_URL	URL of your Ollama instance	http://ollama:11434
OLLAMA_MODEL_ID	The ID of the Ollama model to use	cogito:14b
OLLAMA_MODEL_TEMPERATURE	The temperature for the AI model	0.65
OLLAMA_MODEL_CONTEXT_LENGTH	The context length for the AI model	4096
DEBUG_MODE	Enable or disable debug mode	false
SEARXNG_SECRET	A secret key for SearXNG	your-secret-key

🧑‍💻 Development

To run Aria in development mode, you can use the following command:

python -m aria.cli run --reload

This will start the FastAPI server with hot-reloading enabled.

📄 License

This project is licensed under the MIT License. See the LICENSE file for details.

📚 Documentation

Comprehensive documentation is available on our GitHub Pages site:

• Aria Documentation

The documentation includes:

• Installation Guide
• Configuration Options
• Usage Guide
• Advanced Features
• License Information
• Contributing Guidelines