Atom8 is a state-of-the-art visual AI platform designed to transform messy, unstructured multimedia data into clean, structured datasets. By leveraging high-performance LLMs (Groq & Gemini), it provides a seamless pipeline for data extraction, normalization, and forensic analysis.
Atom8 automates the tedious process of manual data entry and cleaning. Whether you have scanned invoices, messy CSVs, or unstructured web content, Atom8's orchestrator cleans, merges, and validates the data, delivering it in JSON or CSV formats along with a comprehensive "Data Autopsy" PDF report.
- Visual AI Extraction: Directly extract structured information from multimedia and unstructured sources.
- Micro-Orchestration Pipeline: Multi-stage processing including Extraction, Merging, Noise Removal, and AI Structuring.
- Direct CSV Conversion: High-efficiency path that bypasses JSON intermediate for faster CSV processing.
- Data Autopsy Reports: Professional PDF reports generated by AI that analyze data health, error patterns, and AI readiness.
- Multi-LLM Support: Seamlessly switch between Groq (Llama 3) and Google Gemini.
- Real-time Monitoring: WebSocket-based progress tracking from ingestion to export.
Atom8 uses a modern decoupled architecture:
- Frontend: A sleek, reactive dashboard built with Next.js 15, Tailwind CSS, and Framer Motion. It uses Zustand for lightweight state management and provides a premium "Glassmorphism" aesthetic.
- Backend: A robust FastAPI server that manages long-running AI orchestration tasks. It utilizes BackgroundTasks for non-blocking processing and WebSockets for real-time UI updates.
graph LR
User -->|Upload files/URLs| Frontend
Frontend -->|POST Request| Backend
Backend -->|Background Task| Orchestrator
Orchestrator -->|Ingestion| LLM_Service
LLM_Service -->|AI Structuring| Groq/Gemini
Orchestrator -->|Cleaning/Insights| Validator
Orchestrator -->|Export| PDF_Generator
Backend -->|WebSocket Updates| Frontend
- Framework: Next.js 15 (App Router)
- Language: TypeScript
- Styling: Tailwind CSS, PostCSS
- Animations: Framer Motion, Motion
- State: Zustand
- Icons: Lucide React
- Framework: FastAPI
- LLM Engine: Groq (Llama 3.3 70B), Gemini 2.5 Flash Lite
- PDF Core: ReportLab
- Data Processing: Pandas, Unstructured.io
- Package Manager: UV
- Node.js 18+
- Python 3.10+
- API Keys for Groq and/or Google Gemini
-
Clone the repository:
git clone https://github.com/riteshgharat/atom8.git cd atom8 -
Setup Backend: Follow instructions in backend/README.md.
-
Setup Frontend: Follow instructions in frontend/README.md.
Atom8 is developed for hackathon purposes. All rights reserved.