AI Venue Management System — CrewAI Multi-Agent

Project Overview

An AI-powered Venue Management System designed for venue operators (bars, restaurants, hospitality) who want their social presence run by an autonomous pipeline rather than a marketing team. The system uses the CrewAI agentic framework to coordinate three specialised agents that, together, take a venue from "no content" to "scheduled, on-brand social posts" without human intervention in the loop.

The Three Agents

• Content Scraper — pulls source material: menu updates, event posters, photos, supplier notes, and venue-specific signals.
• Content Filter — evaluates scraped material against the venue's brand voice, recency, and compliance rules; rejects what doesn't fit.
• Content Creator — generates post copy, AI images, and short promotional videos, then publishes them through the Facebook and Instagram Graph APIs.

Key Capabilities

• End-to-end autonomy: from raw input → vetted brief → finished post → published asset.
• Multi-modal generation: copy + AI image + short-form video pipeline, owned entirely by the Content Creator agent.
• Direct platform integration: posts go live via Facebook Graph API and Instagram Graph API — no manual export/upload step.
• Per-venue brand voice: filter rules and tone profiles are venue-scoped, so multi-venue operators don't get a generic feed.

Technical Implementation

• Agentic framework: CrewAI (Python) — agent definitions, task orchestration, and tools
• LLM: OpenAI models for copy generation and image prompting
• Publishing: Facebook Graph API, Instagram Graph API
• App surface: Next.js 14 + TypeScript + Supabase (PostgreSQL) for venue/account management, real-time subscriptions, and row-level security

Outcome

A working pipeline that lets a single venue operator maintain a multi-platform social presence without staffing it — the agents do the scraping, the curation, and the publishing.