πŸ”¬

Copernicus AI

A Knowledge Engine for Scientific Discovery

Transforming cutting-edge research into accessible AI-powered podcasts and building a comprehensive knowledge graph connecting concepts, papers, and insights across scientific disciplines.

🎯 Our Mission

Inspired by the historical Copernicus who challenged accepted knowledge with evidence and rigorous analysis, Copernicus AI aims to expand human knowledge by creating tools that help us verify, generate, and transmit scientific insights. As AI systems become more capable, we build instrumentsβ€”like the Programming Framework, GLMP, and AI podcastsβ€”that enable humans to explore, understand, and communicate knowledge with the same depth and speed that AI possesses.

πŸš€ Core Components

πŸŽ™οΈ

AI Podcast Generation

Multi-voice AI podcasts (5-10 minutes) synthesizing research papers into engaging conversations. Powered by Google Gemini 2.0 for content and ElevenLabs for realistic voice synthesis.

  • βœ“ 32+ episodes across 5 disciplines
  • βœ“ Scientific references included
  • βœ“ Subscriber-generated content
  • βœ“ RSS feed distribution
πŸ”§

Programming Framework

A meta-tool for process analysis combining LLMs with Mermaid visualization. Universal method for dissecting complex processes across any discipline.

  • βœ“ Domain-agnostic analysis
  • βœ“ Visual flowchart generation
  • βœ“ Process decomposition
  • βœ“ Knowledge extraction
🧬

GLMP - Genome Logic

Genome Logic Modeling Project: A specialized "microscope" applying the Programming Framework to visualize biochemical processes as interactive flowcharts.

  • βœ“ Biological process mapping
  • βœ“ JSON-based flowcharts
  • βœ“ Pathway visualization
  • βœ“ Research insights
πŸ“š

Research Papers Database

Centralized repository of scientific literature with AI preprocessing. Extracts key findings, entities, and summaries for reuse across podcasts.

  • βœ“ DOI/arXiv integration
  • βœ“ LLM-powered preprocessing
  • βœ“ Entity extraction (genes, proteins)
  • βœ“ Citation tracking
πŸ•ΈοΈ

Knowledge Graph (Phase 2)

Connects concepts, papers, podcasts, and GLMP visualizations. Enables discovery of cross-disciplinary patterns and semantic search.

  • βœ“ Concept relationships
  • βœ“ Paper-podcast links
  • βœ“ Cross-discipline discovery
  • βœ“ Semantic queries
πŸ‘₯

Subscriber Platform

YouTube-style interface for creators to prompt, manage, and share podcasts. Browse public catalog or generate custom content.

  • βœ“ Account management
  • βœ“ Custom podcast generation
  • βœ“ Source paper integration
  • βœ“ RSS feed publishing

πŸ—„οΈ Database Architecture

Firestore Collections

  • subscribers - User accounts, preferences, usage analytics
  • podcast_jobs - Generated podcasts with metadata, engagement metrics
  • research_papers - Scientific papers with AI preprocessing

Google Cloud Storage

  • audio/ - MP3 podcast files (multi-voice synthesis)
  • transcripts/ - Full text transcripts
  • descriptions/ - Markdown descriptions with references
  • thumbnails/ - AI-generated episode artwork
  • glmp-v2/ - Genome Logic flowcharts (JSON)

Enhanced Metadata (Phase 1)

Each podcast now includes: source papers (DOIs/URLs), extracted keywords, quality scores, GLMP visualization links, and engagement metrics (plays, ratings, shares).

πŸ“Š Scientific Disciplines

🧬 Biology (9 podcasts)
βš—οΈ Chemistry (4 podcasts)
πŸ’» Computer Science (6 podcasts)
πŸ“ Mathematics (6 podcasts)
βš›οΈ Physics (7 podcasts)

βš™οΈ Technology Stack

AI & ML

  • β€’ Google Gemini 2.0 Flash
  • β€’ Vertex AI
  • β€’ ElevenLabs TTS
  • β€’ Entity extraction

Backend

  • β€’ FastAPI (Python)
  • β€’ Google Cloud Run
  • β€’ Firestore (NoSQL)
  • β€’ Cloud Storage

Frontend

  • β€’ Static HTML/Alpine.js
  • β€’ Tailwind CSS
  • β€’ Vercel (hosting)
  • β€’ Responsive design

πŸ”— Links & Resources

πŸ”Œ API Endpoints

Base URL: https://copernicus-podcast-api-phzp4ie2sq-uc.a.run.app

Podcast Generation

  • POST /generate-podcast-with-subscriber
  • GET /api/subscribers/podcasts/{id}
  • POST /api/subscribers/podcasts/submit-to-rss

Research Papers (Phase 1)

  • POST /api/papers/upload
  • GET /api/papers/{paper_id}
  • POST /api/papers/query
  • POST /api/papers/{id}/link-podcast/{id}

🌟 The Vision

"I want to know a lot, and I want to be confident that I am learning and disseminating the truth."

As LLMs and other forms of AI gain more knowledge and intelligence, the Copernicus AI project enables humans to keep up with what the AIs are capable of knowing and revealing. We're building the infrastructure for human-AI collaborative knowledge exploration, with truth-seeking as our North Star.