College WhatsApp AI Receptionist with OpenAI & Supabase (n8n)

Turn your college admissions WhatsApp into an AI receptionist—powered by OpenAI + Supabase (built for n8n)

This n8n workflow ingests your college documents into a Supabase vector store and then runs a WhatsApp-based admissions receptionist that can answer text, voice notes, and image enquiries. It retrieves relevant information, reviews the response with a second OpenAI step, logs everything in Supabase, and sends the final reply back to the student via WhatsApp.

What this workflow does

• Receives a WhatsApp message and detects whether the user sent text, a voice note, or an image.
• Transcribes voice notes: fetches the WhatsApp media URL, downloads the audio, and uses OpenAI to convert speech into text.
• Understands images: fetches the media URL, downloads the image, and uses OpenAI Vision to extract readable content (combined with any image caption).
• Normalizes input into a single “question” format, then creates a conversation record in Supabase.
• Generates a draft reply with retrieval: an OpenAI chat agent uses the Supabase vector-store retriever to find relevant college information (while chat history is stored in Postgres).
• Reviews and validates the answer: a second OpenAI reviewer checks the draft against the same Supabase knowledge base—then approves it, retries once with correction instructions, or falls back to a safe response.
• Sends the final WhatsApp reply and updates the Supabase conversation record.

Use cases

• Prospective students asking admission FAQs on WhatsApp (documents like brochures, eligibility notes, or course details pre-indexed).
• Students sending voice notes for guidance; the workflow transcribes, retrieves, and responds.
• Students uploading screenshots of requirements or forms; the workflow extracts text via Vision and answers using retrieval.

Technical details

• n8n nodes/workflow components: if, set, code, switch, Supabase, WhatsApp.
• Integrations: WhatsApp Cloud API (WhatsApp trigger + WhatsApp nodes), OpenAI (chat, transcription, Vision), Supabase vector store (retrieval + Postgres chat history).
• Core logic: document ingestion → vector indexing → multimodal input handling → retrieval-augmented response → OpenAI reviewer step → Supabase logging → WhatsApp delivery.