A custom-built, Java-based AI assistant where I’ve integrated chat memory, retrieval-augmented generation (RAG), tool capabilities, and prompt templates.
🔗 Neo4j for chat memory – Keeps track of user sessions & history for contextual conversations
📄 Template-based prompts – Users can summarize, expand, or simplify text in one click
📚 RAG support with vector embeddings – Texts are embedded & stored in Pinecone for similarity-based retrieval
🎧 Audio transcription – Users can convert audio into text with high accuracy
🌐 Web search integration – AutoWiredGPT can retrieve information beyond the model's training cutoff
AutoWiredGPT is capable of supporting multiple users with distinct, persistent chat sessions, maintaining memory across sessions.
🛠️ Tech Stack Overview
- Java
- Spring Boot (MVC, JPA, AI)
- HTML / CSS
- Neo4j ,Pinecone Vector DB, Neon (PostgreSQL)