CorpMindAI is a secure, intelligent corporate knowledge base that allows teams to chat with their documents. By leveraging RAG (Retrieval-Augmented Generation) technology, it provides accurate, context-aware answers based on uploaded internal files (PDF, DOCX, TXT) without hallucinating.
Built with the PERN stack (PostgreSQL, Express, React, Node.js) and Vector Search, fully containerized with Docker.
The entry point to the system, featuring a modern, responsive design.
Manage your projects and see your team status at a glance.
Ask questions about your documents. The AI cites the context it used to generate the answer.
Easily expand your team. Invite colleagues to specific workspaces via email, view current members, and manage access permissions directly from the workspace settings.
Securely upload and index corporate documents for the knowledge base.
- 🤖 RAG-Powered Chat: Uses Google Gemini Flash combined with ChromaDB to answer questions strictly based on your uploaded data.
- 📂 Multi-Format Support: Upload and index
.pdf,.docx, and.txtfiles automatically. - 🏢 Workspaces & Teams: Create isolated workspaces and invite team members via email.
- 📱 Fully Responsive: Adaptive UI that works seamlessly on Desktop, Tablet, and Mobile.
- 🔐 Role-Based Access: Granular permissions (
OwnervsMember) for workspace management. - ⚡ Real-time Vectorization: Local embedding generation using
@xenova/transformers(runs on CPU, no external embedding API cost). - 🐳 Fully Dockerized: One command to set up the Frontend, Backend, Database, and Vector DB.
- React 18 (Vite)
- TailwindCSS (Custom Dark/Gold Theme)
- Lucide React (Icons)
- Axios (API Communication)
- Node.js & Express
- Multer (File Handling)
- Pdf-parse / Mammoth (Text Extraction)
- PostgreSQL 16 (Users, Chats, Metadata)
- ChromaDB (Vector Database for Embeddings)
- Google Gemini API (LLM for Generation)
- Xenova Transformers (Local Embeddings)
- Docker & Docker Compose
- Nginx (Reverse Proxy for Client)
- Docker and Docker Desktop installed.
- A Google Gemini API Key (Free tier available).
-
Clone the repository:
git clone https://github.com/OkayMarta/CorpMindAI cd CorpMindAI -
Configure Environment Variables: Create a
.envfile in the root directory:cp .env.example .env
Fill in your details in
.env:# Database Configuration POSTGRES_USER=postgres POSTGRES_PASSWORD=secure_password POSTGRES_DB=corpmind_db # Security JWT_SECRET=random_long_string_for_security # AI Service GEMINI_API_KEY=your_google_gemini_key # Email Service (for invitations) EMAIL_USER=your_email@gmail.com EMAIL_PASS=your_app_password
-
Run with Docker: Start the entire application suite:
docker-compose up --build
-
Access the App: Open your browser and navigate to: 👉 http://localhost
- Upload: User uploads a PDF.
- Processing: Server extracts text and splits it into "chunks" (2000 chars).
- Embedding: Local AI model converts chunks into numerical vectors.
- Storage: Vectors are stored in ChromaDB, metadata in PostgreSQL.
- Chatting:
- User asks a question.
- System converts question to vector.
- System finds the top 15 most similar text chunks in ChromaDB.
- System sends chunks + question to Google Gemini.
- AI generates an answer based only on those chunks.
This project is licensed under the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0).
You are free to:
- Share — copy and redistribute the material in any medium or format.
- Adapt — remix, transform, and build upon the material.
Under the following terms:
- NonCommercial — You may not use the material for commercial purposes (selling the app, SaaS, providing paid services using this code) without explicit permission.
- Attribution — You must give appropriate credit.
For commercial inquiries or licensing, please contact the author.