AthenusAthenus is a self-hosted AI learning companion. Transform video lectures into interactive learning workspace made for you.
# Clone the repository
git clone https://github.com/Famanias/athenus.git
cd athenus
# Create .env from the template (required)
cp .env.example .env
# Start the full CPU stack (backend :8000, frontend :47734, Ollama :11434)
docker compose up -d --build
# Open http://localhost:47734 in your browser.
DISCLAIMER: 99% of the code is AI Generated. The other 1% is me debugging and trying to figure out which architecture and tech stack to use.
Dedicated for Learning
Everything you need to master video lectures, course series, and complex educational media in one unified application.
Import Lectures
Drag and drop any lecture. Athenus automatically transcribes and prepares it for learning.
AI Chat
Ask questions in natural language and get answers backed by your lecture with clickable citations.
Built for Learning
Search transcripts, revisit key moments, and explore concepts. All in one place.
Import your Own
Athenus uses local LLM models. But you can also bring your own API key.
Workspaces & Multi-Session Organization
Keep every course, topic, or project in its own workspace.
Local-First Knowledge Storage
Everything stays on your computer. No stealing or selling your information.
Persistent Memory
Your settings, workspaces, and conversations are saved automatically, and you stay in control.
Open Source & Free
MIT Licensed. Clone it, fork it, or steal it (please don't).
Why Athenus?
Athenus started as a weekend project inspired by PewDiePie's Odysseus. I just wanted something that could help me understand the online courses I was taking. Turns out there was more to this, I figured if it could help me learn, it might help someone else too.
Also the name Athenus is just a masculine-version of Athena. I simply thought all the Athena domains are taken due to popularity.
The Roadmap
Athenus is evolving from a local-first RAG application into a local-first AI knowledge and learning operating system. V1 — video and document ingestion, the Knowledge Graph, Flashcards, Quizzes, and the Mastery dashboard — is shipped. The next chapter is Foundation: making every answer trustworthy, every citation a first-class object, and the whole knowledge base reachable to other AI systems through MCP.
Local-First AI Learning Space
The shipped foundation: ingestion, knowledge graph, recall, and mastery.
V1 is the complete first chapter of Athenus. Every video, document, and concept inside a workspace is searchable, citable, and recallable — and the entire pipeline runs locally on your own hardware.
- Local-first video & document ingestion with full RAG
- Interactive Knowledge Graph of extracted concepts
- AI Flashcards & adaptive Quizzes with spaced repetition
- Analytics dashboard with concept mastery & retention
Foundation
Make retrieval trustworthy before everything else.
Before Athenus grows new capabilities, retrieval and generation have to be defensible. Foundation turns the RAG pipeline into a system you can interrogate and audit, and makes citations a first-class output of every answer.
- RAG Reliability — hybrid retrieval, reranking, confidence scoring, grounded generation
- First-Class Citation System — structured citations that jump to exact pages or timestamps
- Universal Search (Ctrl+K) across every artifact in the workspace
- Athenus MCP Server — expose the knowledge base to other AI systems
Intelligence
Agents, memory, and a knowledge layer that actually understands.
Intelligence is where Athenus stops being a knowledge store and starts being a collaborator. Specialized agents operate against your real workspace, reusable skills define how they think, and personal memory lets Athenus adapt explanations to you.
- Specialized Agents — Study, Research, Lecture, Exam Prep, Writing, Coding
- Reusable Skills — Explain Like I'm 5, Cornell Notes, Summarize Lecture, Find Knowledge Gaps
- Knowledge Graph as an entity layer — concepts, people, topics, formulas, projects
- Personal Memory — preferences, learning goals, weak concepts, prior questions
- Deep Research pipelines combining web sources + your workspace
- Model Management — providers, context windows, capabilities, cost
- Learning Intelligence — connect recall to a broader mastery loop
Productivity
Workflows, notes, voice, and scheduled learning.
Productivity wraps Athenus around how you actually learn. Notes stay connected to knowledge entities, workflows automate the boring parts, voice interaction rides on top of existing transcription, and scheduled learning ties spaced repetition to your calendar.
- Integrated Notes — deeply linked to Athenus knowledge entities
- AI Workflows — trigger / AI action / condition chains
- Library / Artifacts — central hub for user- and AI-generated outputs
- Voice Interaction — voice questions & voice notes layered on transcription
- Scheduled Learning — Monday review, weekly reports, post-upload analysis
- Model Comparison — same prompt across A/B/C models with groundedness scoring
Future Expansion
Beyond the desktop app.
Long-term expansion that takes Athenus out of a single device and into a broader ecosystem, without compromising the local-first identity that defines it.
- Mobile companion app
- Multi-user collaboration
- External integrations
- Plugin ecosystem
- Third-party MCP ecosystem
- Cross-device synchronization
Support the Open Source Mission
Athenus is and always will be open source and free forever.
Tips are completely optional.