Development of an Intelligent Multichannel Ai Assistant with Dynamic Knowledge Base

We are seeking an experienced AI developer or team to create an advanced Intelligent Multichannel Reception Assistant. This assistant will be capable of receiving and processing queries and data from various sources, primarily WhatsApp, and handling information in diverse formats:

*  Images: Implement visual analysis capabilities for extracting relevant data from images.
*  Audio: Develop precise transcription of voice messages to text, utilizing advanced technologies such as ElevenLabs to ensure complete information capture.
*  Text: Enable direct and efficient processing of written messages.

Once converted to text, the data will undergo advanced Natural Language Processing (NLP) using high-quality embeddings. This is crucial for achieving a deep understanding of the intent and meaning behind each user query.

The core of this project is the integration with a Dynamic Knowledge Base, powered by Retrieval-Augmented Generation (RAG). This will involve setting up and integrating with a vector database in Supabase. The RAG system will ensure that the assistant's responses are highly accurate, contextually relevant, and strictly based on real company data, thereby preventing 'hallucinations' and guaranteeing reliability. All interactions must be recorded in an analyzable database for future insights and improvements.

Finally, the agent will operate under a strict 'Model Context Protocol' to generate intelligent, detailed, and personalized responses using advanced language models. The goal is to provide a seamless, accurate, and highly effective communication tool for our users.