This internal tool enables AI-driven document curation for NASA teams. It provides a secure web-based interface to upload scientific files, generate AI summaries and keywords, and interact with content through a Retrieval-Augmented Generation (RAG) chat interface.
🔒 Note: A version of this tool is hosted on NASA’s SMCE and integrated with the Open Science Data Repository (OSDR). That instance is accessible only via NASA VPN and intended for internal use by the OSDR curation team.
- Upload
.pdf,.xlsx, and image files - Automatic AI-generated summaries and keyword tagging
- RAG-based chatbot to query your documents in natural language
- Public OSDR API integration for dataset discovery
- Session-based directory isolation for each user
- Works locally or in secure internal environments (e.g., SMCE)
See the Quick Start Guide for step-by-step usage instructions.
This is a Next.js project bootstrapped with create-next-app.
npm run dev
# or
yarn dev
# or
pnpm dev
# or
bun devThen open http://localhost:3000 in your browser.
You can begin editing the page by modifying app/page.tsx. The app will auto-refresh with updates.
This project uses next/font with Geist for optimized font loading.
User (VPN)
↓
Next.js Frontend (React)
↓
Python Backend (FastAPI or Flask)
├── LLM Summarization
├── RAG Chat via local Ollama or HF model
├── OSDR API Integration
└── Session-Based File & Embedding Storage
- Hosted on NASA SMCE (internal deployment)
- Session directories isolated per user
- Tested iteratively during development by the internal project team
- Issues tracked via GitHub Projects, categorized by priority
- 508 accessibility compliance verified through manual review and ticketing
- Open source and adaptable for other NASA applications
Have ideas or want to report an issue?
- Submit tickets via GitHub Issues
- Use the internal project board for task tracking
- Contact the dev team directly on Slack or via email
- Quick Start Guide
- OSDR Reference: https://visualization.osdr.nasa.gov/