Company Knowledge Base RAG Agent with Auto-Sync

This workflow transforms your company documents into an intelligent AI assistant using Retrieval-Augmented Generation (RAG). Triggered by Telegram messages (easily adaptable to websites), it handles both text and voice inputs—transcribing audio via OpenAI Whisper—then queries a Supabase vector database for relevant document chunks using OpenAI embeddings. An LLM generates context-aware responses, reducing repetitive support queries and empowering teams with instant access to internal knowledge.

Two companion workflows ensure your knowledge base stays current: the 'New File Workflow' detects uploads in a Google Drive folder, extracts text, splits it into chunks, generates embeddings, and upserts them into Supabase. The 'Update File Workflow' monitors changes, removes outdated entries, and refreshes embeddings for edited files, automating maintenance entirely.

Ideal for startups building HR/ops assistants, SaaS helpdesks, customer support, or knowledge teams. Benefits include 24/7 query handling, reduced ticket volume, scalable doc management, and seamless integration—saving hours weekly on manual searches and updates. Deploy as internal tool or public bot for maximum impact.