Citation Sentinel
A source-grounded research assistant inspired by Google's NotebookLM. MIT license, by @sjdev. Users upload documents, links, youtube videos, ask questions and receive answers (with inline citations) grounded in their sources.
Built with a React/Vite frontend and a Node.js/Express backend, using Anthropic Claude for generation, OpenAI Whisper for video audio track transcription, Voyage AI for embeddings and response cosine similarity scoring ("groundedness" score).
Prerequisites
- Node.js (v18+)
- yt-dlp -- required for YouTube video source support (
brew install yt-dlporpip install yt-dlp)
Getting Started
# clone and install
git clone <repo-url> && cd notebooklm_clone
cd server && npm install && cd ..
cd client && npm install && cd ..
# configure
cp server/.env.example server/.env
# edit server/.env and add your ANTHROPIC_API_KEY, VOYAGE_API_KEY,
# and optionally OPENAI_API_KEY (only needed for audio source transcription via Whisper)
# run
make dev
Rationale
NotebookLM is a powerful research tool, but it is proprietary and closed. This clone demonstrates the core source-grounded Q&A pattern with transparent retrieval, generation, and groundedness scoring (LLM response quality cosine scoring) -- all with swappable models/fully open source.
Design
Two-package monorepo:
-
server/-- single Express backend with layered architecture (routes -> services -> stores). Routes orchestrate; services contain business logic; stores manage in-memory state. -
client/-- React 19 SPA via Vite. Two-panel layout: sidebar for notebooks/data sources, main area for LLM chat with explorable citations, groundedness badges (cosine similarity scoring of LLM responses), and follow-up question chips.
Query pipeline
Embed query (Voyage) -> k-NN search -> rerank (Voyage) -> generate answer with citations (Claude) -> compute groundedness score (cosine similarity of answer vs cited chunks).
License
MIT. See LICENSE.