
Curated by n8n Markets team
Last updated: March 8, 2026
### π§ Intelligent AI Assistant with RAG & Voice for WhatsApp β Built with GPT-4o & Supabase --- ### π About this workflow and its creator Hi! Iβm Amanda, a creator of intelligent automations using n8n and Make. Iβve been building AI-powered workflows for over 2 years, always focused on usability and innovation. This one here is very special to me β a truly advanced AI assistant that reads, listens, interprets and responds like a real human π€β¨ This ready-to-use workflow acts as a powerful AI personal assistant capable of understanding messages via voice, text, documents, or even images. It supports full multi-channel operation (WhatsApp via Evolution API, Instagram, Facebook, and more), and includes advanced RAG capabilities using Supabase + GPT-4o. Itβs designed to be highly extensible, with memory, prompt update tools, and knowledge base management. --- ### βοΈ What this workflow does - π¬ Understands user input via text, document, audio or image (voice, OCR, PDF) - π€ Transcribes and interprets voice messages using OpenAI Whisper - π§ Understands prompts and user commands using GPT-4o via LangChain agent - ποΈ Searches knowledge base using RAG + Supabase vector DB - π Accepts documents and automatically indexes them for future questions - π§Ύ Summarizes documents and stores metadata in Supabase - ποΈ Offers memory support (PostgreSQL chat memory per user session) - π§ Sends replies through WhatsApp (Evolution API), Instagram, Facebook, etc. - π Manages schedules (via tool integration with Google Calendar) - π¬ Sends and searches emails (with support tools) - π Modular and expandable structure (tools for saving knowledge, deleting, updating prompt) --- ### π§ Setup Instructions 1. n8n Hosting This workflow requires n8n self-hosted (or n8n Cloud with custom credentials + community nodes enabled). 2. Create required databases Use the provided SQL queries inside the `setar_supabase_tabelas_vectoriais`, `criar_cerebro`, and `criar_rag_controle` nodes to initialize: - `documents` table for RAG - `cerebro` table for prompt - `memoria_chat` for session memory - `rag_controle` for summaries and indexing 3. Credentials needed - OpenAI API (for chat, embeddings and Whisper transcription) - Redis (for managing message buffer) - Supabase (for vector store + metadata) - Postgres (for memory and prompts) - Evolution API (or other messaging platforms) 4. Webhook Set the webhook path to receive messages from your Evolution or WhatsApp API provider. 5. Configure βSetβ node In the `config` node, adjust: - `adminNumero`: your personal WhatsApp or admin number - `evolutionApiKey`: your private API key - `utilizacaoApenasViaAdmin`: toggle if this should only respond to admin numbers 6. Tool connections Ensure the supporting workflows are also imported and connected for: - Emails - Knowledge management - Calendar events --- ### π Notes - This workflow uses LangChain agents, OpenAI GPT-4o, Supabase, Redis, and PostgreSQL. - It includes multiple βsticky notesβ inside the workflow with explanations. - Ideal for businesses, consultants, and developers looking to offer an intelligent and extendable AI chatbot experience. --- ## π Want to use this on your system? β€οΈ Buy workflows: [https://iloveflows.com](https://iloveflows.com) βοΈ Use n8n Cloud with my partner link: [https://n8n.partnerlinks.io/amanda](https://n8n.partnerlinks.io/amanda)
n8n instance (self-hosted or cloud). credentials.
1. Download the workflow JSON 2. Import into your n8n instance 3. Configure credentials for 4. Activate and enjoy automation
Included in Builder Pro β $19/mo, 4,177 premium workflows
Templates are reviewed before they are listed. Test with your own credentials and data before running it in production.
100% Secure Checkout
Refund requests within 7 days of a new subscription
0 reviews
No reviews yet.
Log in to ask a question.
No questions yet. Be the first to ask!