SciTeX

What You Get

SciTeX Cloud is a self-hostable research platform. One make start gives you a full Django application with:

Scholar — Literature search across CrossRef, PubMed, arXiv, and OpenAlex. BibTeX management and citation tracking.
Writer — LaTeX manuscript editor with BibTeX integration, figure/table management, and IMRAD templates.
Vis — Data visualization and interactive figure editing.
Console — Web-based terminal for running Python and Bash in isolated containers.
Hub — Project file browser with GitHub-style /username/project/ URLs.
Research Tools — Statistics, PDF manipulation, citation scraping, QR generation, and more at /tools/.
MCP Server — 23 tools for AI agents (Claude, Cursor, etc.) to search literature, manage citations, generate figures, and run statistics.

Ecosystem Overview

Architecture

Docker containers, fully orchestrated:

Django (port 8000) — Web application and REST API
PostgreSQL — Database for all application data
Redis — Cache and Celery message broker
Gitea (port 3000) — Self-hosted Git server for project repositories
Celery — Async task workers for background jobs
Umami (port 3300) — Privacy-focused analytics (no external tracking)

Optional local databases for offline paper search:

CrossRef Local — 167M+ papers in a local SQLite database. Fast, offline metadata search with citation graph analysis.
OpenAlex Local — 284M+ scholarly works with full-text search, abstracts, and impact factors.

Infrastructure Deployment

Complete infrastructure stack showing Docker services, host services, and network architecture.

Quick Start

Three commands to get everything running:

git clone https://github.com/ywatanabe1989/scitex-cloud.git
cd scitex-cloud
make start

Access at localhost:8000. Test user: test-user / Password123!

Prerequisites

Linux, macOS, or Windows (WSL2)
Docker 24.0+ with Docker Compose v2
Git 2.30+
4 GB RAM minimum (8 GB recommended)

Development Setup

Clone and configure

git clone https://github.com/ywatanabe1989/scitex-cloud.git
cd scitex-cloud
cp deployment/docker/envs/.env.example deployment/docker/envs/.env.dev

Set secret key (edit .env.dev)

SCITEX_CLOUD_DJANGO_SECRET_KEY=$(python3 -c 'import secrets; print(secrets.token_urlsafe(32))')
SCITEX_CLOUD_POSTGRES_PASSWORD=your-dev-password

Start services
```
make start
make status
```

Production Setup

For deploying on a home server, NAS, or VPS:

Create production environment

cp deployment/docker/envs/.env.example deployment/docker/envs/.env.prod

Configure for production (edit .env.prod)

DEBUG=False
SCITEX_CLOUD_DOMAIN=scitex.example.com
SCITEX_CLOUD_SITE_URL=https://scitex.example.com
SCITEX_CLOUD_ENABLE_SSL_REDIRECT=true

Start production

make ENV=prod start
make ENV=prod status

CLI-Only Install

Use the CLI and MCP server without Docker:

pip install scitex-cloud              # CLI only
pip install scitex-cloud[mcp]         # CLI + MCP server
pip install scitex-cloud[all]         # Everything

scitex-cloud --version
scitex-cloud --help

MCP Server for AI Agents

Connect AI assistants (Claude, Cursor, etc.) to your SciTeX instance:

scitex-cloud mcp start                # Start MCP server
scitex-cloud mcp doctor               # Diagnose setup
scitex-cloud mcp installation          # Client config

Common Commands

make start                    # Start development
make stop                     # Stop all services
make restart                  # Restart services
make status                   # Health check
make logs                     # View all logs
make ENV=prod start           # Start production
make help                     # All available commands

Full Documentation

For configuration reference, troubleshooting, and architecture details, see the full setup guide on Read the Docs.