Empower your enterprise with a customizable chatbot service that leverages your internal knowledge base to provide accurate, context-aware answers.

Chat-Craft is a chatbot-as-a-service RAG platform built for enterprises seeking to break down knowledge silos and streamline internal information retrieval. By leveraging your organization's documents and data, Chat-Craft delivers reliable, context-aware answers using advanced Retrieval Augmented Generation (RAG) techniques.

• Document ingestion & vectorization: Seamlessly upload and process internal documents for instant Q&A capabilities.
• Vector search for relevant context: Retrieve the most pertinent information from your knowledge base for every query.
• Structured responses via LLMs: Ensure consistent, accurate answers with large language models fine-tuned for your data.
• Microservices architecture: Robust, scalable, and maintainable system design.
• Secure & private deployment: Keep your data safe with enterprise-grade security and deployment options.
• Customizable UI components: Tailor the chatbot experience to your brand and workflow.
• Exportable iFrame Widget: Easily embed your chatbot in any internal tool or website.

Chat-Craft employs a modern microservices architecture, orchestrated by a central service that manages document processing, vector storage, and AI-powered responses.

Core Services:

• Entities Service (Orchestrator)
• File AI Service
• Cron Job Service
• UI Frontend

Technology Stack:

• Next.js, Tailwind CSS, Interactive UI Components

• FastAPI for backend services

• MongoDB with Atlas Vector Search

• Ollama (local LLMs), Nomic Embed Text (embeddings), Llama 2 (response generation)

  

1. User uploads documents
2. File AI Service processes and splits documents into chunks
3. Nomic Embed Text converts chunks into vector embeddings
4. Embeddings stored in MongoDB with user metadata

1. User submits a query via the chat interface
2. Query is embedded into a vector
3. MongoDB Atlas Vector Search retrieves relevant document chunks
4. Chunks are sent to Llama 2 with a structured prompt
5. Llama 2 generates a context-aware, structured response
6. Response is displayed in the chat interface

• Dashboard: Manage your chatbot, analyze performance, and train with your documents.
• Chatbot UI: Interactive, user-friendly chat interface for seamless Q&A.
• Configuration & Analytics: Customize the system and gain insights into usage and knowledge gaps.
• Exportable Widget: Integrate the chatbot in your internal tools with a simple iFrame.

• Internal IT or HR support
• Onboarding and training assistance
• Technical documentation Q&A
• Customer support knowledge base

• Empower your teams: Fast, accurate answers reduce time spent searching for information.
• Break down silos: Centralize knowledge and make it accessible to everyone.
• Customizable & secure: Tailor the chatbot to your needs and keep your data private.

Deploy Chat-Craft in your organization and transform how your teams access knowledge!