AI Community

On this page

• Overview & Vision
• Platform Capabilities
• Quick Start
• Setup Wizard
• Core User Areas
• Admin Operations
• Architecture
• Configuration & Deployment
• Database Support
• API & Integration
• MLStudio & Ecosystem Integration
• Developer Guide
• In-App Help
• Troubleshooting & FAQ

Overview & Vision

AI Community (technical/package name Beep.AI.Community) is a Flask web application that makes machine learning accessible to everyone — students, business professionals, and curious explorers alike, with no technical background required. Users can share data with simple uploads, browse and learn from others' work, join friendly competitions with guided steps, and get community support built into the platform.

The application is designed to be branded and deployed per organization: it supports setup-time company identity and industry themes (for example Oil & Gas, Finance, or a General theme), multiple database backends, and two authentication models. It is part of the Beep ecosystem and integrates with MLStudio for one-click project publishing.

Platform Capabilities

Quick Start

The platform separates installation (handled by the launcher scripts) from configuration (handled by the Setup Wizard on first run). The launchers bootstrap Python, create a virtual environment (.venv), install everything from requirements.txt, prepare directories and a .env file, then start the app. Python 3.8 or higher is required.

run.bat
# Or simply double-click run.bat in Windows Explorer

./run.sh
# On macOS you can also double-click run.command

python run_community.py

# Optional command-line flags:
python run_community.py --port=5003      # Custom port
python run_community.py --host=0.0.0.0   # Listen on all interfaces
python run_community.py --debug          # Enable debug mode
python run_community.py --no-browser     # Do not open a browser automatically

By default the application runs on http://127.0.0.1:5002. On first launch you are automatically redirected to the Setup Wizard; there is no separate setup command or flag.

Setup Wizard (First Run Only)

The wizard handles application-level configuration only — it does not install Python, packages, virtual environments, or system database servers. It walks through the following steps:

Log in with the admin username and password you created during setup.

Core User Areas

The platform is organized around a few high-value workflows:

Admin Operations

Administrators are responsible for setup, branding, authentication, database settings, and ongoing governance:

• Setup & governance — complete the wizard with deliberate auth and database choices; document ownership of admin accounts and support paths; review provider-specific requirements before rollout.
• Branding & identity — set company identity, logo, and industry theme consistently; keep the support email and help links current; validate icon/theme coverage after branding updates.
• Maintenance checklist — review competition and submission health regularly, audit dataset and project quality signals, and keep help/README content aligned with actual platform behavior.

Architecture

The application follows a Flask application factory pattern (create_app) with a clear separation between web routes, REST API routes, the first-run setup flow, services, and models.

Supporting technologies include SQLAlchemy with Flask-SQLAlchemy for portable persistence (tables are created with db.create_all() and Flask-Migrate/Alembic manages migrations), Flask-RESTX for the documented API, Flask-JWT-Extended for token/cookie auth, Flask-SocketIO for real-time updates, Flask-Limiter for rate limiting, Flask-CORS, and Bleach for HTML sanitization. Pandas, NumPy, and Plotly support data processing and preview; Celery with Redis is available for background tasks. Baseline security headers (X-Frame-Options, X-Content-Type-Options, Referrer-Policy, and a Content-Security-Policy) are applied to every response, and a same-origin check mitigates CSRF on cookie-authenticated writes.

Configuration & Deployment

Installation and configuration are intentionally separated:

• Launcher scripts bootstrap an embedded Python runtime when needed (used only for bootstrap), create the .venv application runtime, install dependencies, create directories, and start the app.
• The setup wizard configures the application: admin creation, database provider, authentication mode, and branding/industry setup.

Configuration is portable and repository-local (managed through the config manager and a .env file). Deployment notes: SQLite works for simple or local installations and needs no separate server; PostgreSQL, MySQL/MariaDB, SQL Server, Oracle, and Firebird require their own reachable database server. Identity Server mode requires valid external configuration values (authority URL, client id, and secret).

Database Support

All database drivers ship in requirements.txt and are installed by the launchers; the administrator selects one in the setup wizard.

Persistence uses the SQLAlchemy ORM, so the same models work across every supported provider.

API & Integration

The platform exposes a versioned REST API under /api/v1, built with Flask-RESTX. Interactive Swagger documentation is served at /api/v1/docs. The following namespaces are registered:

Requests are rate-limited by default, and the API accepts JWT tokens via header or cookie. Note that Projects are managed through the web UI rather than a dedicated API namespace.

MLStudio & Ecosystem Integration

AI Community is part of the Beep ecosystem and ships clients for its sibling services:

• MLStudio — MLStudio can publish projects, datasets, and models into the Community platform through the REST API, so work created there can be shared with one click.
• Beep.AI.Server — an optional client for AI-assisted features such as auto-descriptions and LLM-backed help, plus model-related services.
• Identity Server — enterprise deployments can delegate sign-in to Beep.Foundation.IdentityServer via OAuth2/OIDC (authorization-code flow with an /auth/callback redirect handled by the app).

Developer Guide

When extending the platform, keep responsibilities in the right layer:

1. Locate the correct route group before adding an endpoint; put request/response orchestration in routes.
2. Put business rules in services, not in large view functions.
3. Keep templates focused on rendering.
4. Preserve the installation-versus-configuration split.
5. Keep README and in-app help aligned with actual user-facing behavior.

Key modules to understand first: the configuration manager, the setup service, the auth service, the branding service, the main web routes, and the API package under app/routes/api/v1. Branding is treated as part of the platform contract, not a cosmetic afterthought — it flows into navigation, landing pages, and help surfaces.

flask db init          # First time only
flask db migrate -m "Description"
flask db upgrade

In-App Help

Once the application is running, a full help hub is available in the browser:

Troubleshooting & FAQ