Carbon

Open-source research paper management. Import papers, understand them faster, map how they connect, and chat with your library.

Features | Quick Start | Configuration | Database | Project Structure

Overview

Carbon turns arXiv links, PDFs, and research pages into a personal research workspace. It extracts metadata, stores papers in a shared Supabase-backed library, generates AI summaries, builds a similarity graph, and lets you ask grounded questions about each paper.

The app is built with Next.js 16, React 19, Supabase Auth/Postgres/Storage, OpenRouter-hosted LLM calls, Sigma.js graph rendering, and an embedded PDF reader.

Features

• Import arXiv URLs, direct PDF links, or generic paper pages with streaming progress.
• Store a personal paper library with per-user read state, notes, sorting, and filtering.
• Generate structured AI summaries.
• Build paper-to-paper similarity edges with embeddings and pgvector.
• Explore an interactive WebGL knowledge graph powered by Sigma.js and Graphology.
• Read PDFs in-app through a proxy route that avoids common external CORS issues.
• Chat with individual papers using extracted paper text, summary, and library context.
• Use Google OAuth via Supabase, public profiles, follows, stars, claims, and a community activity feed.

Tech Stack

Area	Technology
App	Next.js 16 App Router, React 19, TypeScript
Styling	Tailwind CSS 4, custom global CSS, Next fonts
Auth and data	Supabase Auth, Postgres, Storage, SSR helpers
AI	OpenRouter chat completions and embeddings
Papers	Cheerio, pdf-parse, arXiv scraping, PDF upload/storage
Graph	Sigma.js, Graphology, ForceAtlas2, no-overlap layout
Reading	@embedpdf/react-pdf-viewer, KaTeX, Markdown rendering

Quick Start

Prerequisites

• Bun, Node.js, or another package manager compatible with Next.js.
• A Supabase project with Google OAuth enabled.
• A Supabase Storage bucket named papers.
• An OpenRouter API key.

Install

bun install

Configure Environment

Create a local environment file from the example:

cp .env.example .env.local

Fill in the required values:

NEXT_PUBLIC_SUPABASE_URL=https://your-project.supabase.co
NEXT_PUBLIC_SUPABASE_PUBLISHABLE_KEY=your-publishable-key
SUPABASE_SERVICE_ROLE_KEY=your-service-role-key
OPENROUTER_API_KEY=your-openrouter-api-key

Run Locally

bun dev

Open the app at:

http://localhost:3000

Build

bun run build

Start the production server:

bun start

Usage

1. Sign in with Google.
2. Paste one or more arXiv, PDF, or paper page URLs into the import flow.
3. Select a paper to read metadata, notes, PDF content, and community state.
4. Run enrichment to stream summaries and graph connections.
5. Switch to the graph view to explore related papers.
6. Open the chat tab to ask paper-specific questions grounded in extracted text.

License

MIT