Web Application Installation

This page covers installation and setup of the AIDRIN web interface. For the CLI and the agentic evaluation component, see the CLI Installation page.

AIDRIN can be used in three ways:

  1. Run locally from source — best for development or using the latest GitHub version.

  2. Install as a Python package via PyPI — simplest for using AIDRIN in scripts or notebooks.

  3. Use the hosted web application — no installation required.

Choose the option that best fits your workflow.

Option 1: Local Installation from Source

Works on macOS, Linux, and Windows (via WSL or Anaconda).

Prerequisites

Before installing AIDRIN locally, ensure you have:

Step 1: Clone the Repository

git clone https://github.com/idtlab/AIDRIN.git
cd AIDRIN

Step 2: Set Up the Conda Environment

conda create -n aidrin-env python=3.10 -y
conda activate aidrin-env
python -m pip install -e .

This installs AIDRIN and its dependencies in editable mode.

Optional extras:

# AI-generated explanations of metric results (OpenAI-compatible APIs)
pip install -e ".[llm]"

# Remote metric execution via Globus Compute
pip install -e ".[globus]"

# OpenTelemetry tracing
pip install -e ".[telemetry]"

# All optional features
pip install -e ".[llm,globus,telemetry]"

Step 3: Install Required Services

AIDRIN uses Redis for background task management and Celery for asynchronous execution.

Install Redis Locally

macOS (Homebrew):

brew install redis

Ubuntu/Debian:

sudo apt update
sudo apt install redis-server

Windows:

Verify Redis is running:

redis-cli ping

Expected output: PONG

Step 4: Start the Application

Open three terminal windows/tabs:

Terminal 1 – Redis Server

redis-server --port 6379

Terminal 2 – Celery Worker

macOS / Linux:

conda activate aidrin-env
PYTHONPATH=. celery -A worker.make_celery worker --beat --loglevel=info

Windows:

If you see errors such as:

  • -B option does not work on Windows. Please run celery beat as a separate service.

  • Can't pickle local object 'celery_init_app.<locals>.FlaskTask'

Use the solo pool instead (no --beat required for local development):

conda activate aidrin-env
$env:PYTHONPATH = "."
celery -A worker.make_celery worker --loglevel=info --pool=solo

If you use a venv rather than Conda, activate it first and set PYTHONPATH the same way before running the celery command.

Terminal 3 – Flask Server

conda activate aidrin-env
flask --app 'web:create_app()' run --debug

Note

Windows: If you see -B option does not work on Windows or Can't pickle local object 'celery_init_app.<locals>.FlaskTask', drop --beat from the worker command and use --pool=solo instead:

PYTHONPATH=. celery -A worker.make_celery worker --pool=solo --loglevel=info

To run periodic tasks, start Beat in a separate terminal (optional for local development):

PYTHONPATH=. celery -A worker.make_celery beat --loglevel=info

Once running, visit: http://127.0.0.1:5000

Note

The maximum upload size defaults to 1 GB. To change it for a deployment, set the AIDRIN_MAX_UPLOAD_MB environment variable (in megabytes) before starting the Flask server, e.g. AIDRIN_MAX_UPLOAD_MB=2048 flask --app 'web:create_app()' run.

Option 2: Install from PyPI

For quick use in Python scripts or Jupyter notebooks:

pip install aidrin

To pin a specific version:

pip install aidrin==<version>

Replace <version> with the latest from PyPI versions.

Verify installation:

import aidrin
print(aidrin.__version__)

See Web Application Usage for examples.

Option 3: Use the Hosted Web Application

For zero setup, use the hosted version at: https://aidrin.io

Advantages:

  • No installation or dependencies

  • Runs entirely in the browser

  • Same features as the local version

  • All processing is server-side

Simply upload datasets and run analyses directly from the interface.

Note

Both the local and web versions share the same core codebase. The web version is pre-configured and ready to use, while the local version offers flexibility for customization.