aira - Agentic Information Retrieval Assistant

Aira is a demo AI-powered document chat application that allows users to interact with documents using natural language.

Prerequisites

• Python 3.11+
• uv package manager
• Node.js 18+ (for frontend development)
• npm package manager
• Docker or Docker Desktop (Windows users)
• Make (for managing tasks)

🚀 Quick Start

1. Clone the repository:

git clone https://github.com/philsv/aira.git
cd aira

2. Install dependencies:

make install

3. Environment Configuration

Copy the environment template and configure your services:

cp .env.template .env

Edit the .env file with your service configurations:

Required Environment Variables

OpenAI Configuration (for embeddings):

OPENAI_API_KEY=your_openai_api_key_here
EMBED_MODEL=text-embedding-ada-002
MODEL_DIMENSIONS=1536

xAI Configuration (for chat completions):

XAI_API_KEY=your_xai_api_key_here
XAI_MODEL=grok-3-mini

Qdrant Vector Database:

QDRANT_HOST=host.docker.internal
QDRANT_PORT=6333

AWS S3 Storage:

S3_BUCKET=your_s3_bucket_name
S3_ACCESS_KEY=your_s3_access_key
S3_SECRET_KEY=your_s3_secret_key
S3_REGION=your_s3_region

LangSmith Tracing (optional):

LANGSMITH_API_KEY=your_langsmith_api_key
LANGSMITH_PROJECT=aira-project

Server Configuration:

UVICORN_HOST=0.0.0.0
UVICORN_PORT=8000

4. Deploy the application in docker:

Ensure Docker is running, then build and start the application:

make deploy

🛠️ Development

Start the backend server locally:

make server

Dependency Management

This project uses uv for fast Python package management. Dependencies are defined in requirements.in.

To update dependencies:

make requirements

Testing

Run the test suite:

make test

Accessing the Application

The API will be available at:

• API Server: http://localhost:8000
• Interactive Docs: http://localhost:8000/docs
• ReDoc: http://localhost:8000/redoc
• OpenAPI JSON: http://localhost:8000/openapi.json

To access the chat, run:

• Frontend App: http://localhost:3000

To access the qdrant dashboard, run:

• Qdrant Dashboard: http://localhost:6333/dashboard

To access the fumadocs documentation, run:

• Documentations: http://localhost:3007