SciFlow

English | 中文

Try the SciFlow Online Demo →

Explore the interface instantly with simulated research data—no installation or file upload required.

Start SciFlow locally

Install and start Docker Desktop, then use the launcher for your platform:

macOS	Windows	Linux
Double-click start-local-en.command	Double-click start-local-en.bat	Run ./start-local-en.sh

Open http://127.0.0.1:8000 when startup completes. Your documents and application data stay on your own computer.

SciFlow is a local research document workspace. The React frontend and FastAPI backend are packaged as one local application, while PostgreSQL/pgvector, uploaded files, and the embedding-model cache remain in local Docker volumes. The application has no account system.

English documentation

• User guide: requirements, one-click startup, model configuration, backups, and troubleshooting.
• Developer guide: source development, migrations, API conventions, tests, and packaging.

中文文档：

• 普通用户使用教程
• 开发者二次开发教程

More startup options

Install and start Docker Desktop first.

Platform	English startup	English stop
macOS	Double-click start-local-en.command	Double-click stop-local-en.command
Windows	Double-click start-local-en.bat	Double-click stop-local-en.bat
Linux	./start-local-en.sh	./stop-local-en.sh

Open http://127.0.0.1:8000 after startup.

The generic commands remain available and select Chinese or English from the system locale:

./start-local.sh
./stop-local.sh

You can also use Docker Compose directly:

docker compose up -d --build --wait

Data and privacy

• Documents, database records, and vectors remain in persistent Docker volumes on the user's computer.
• Conversations, favorites, UI settings, and model API keys remain in the current browser.
• The service binds only to 127.0.0.1 and is not exposed to the LAN or internet by default.
• When using a cloud model, questions and retrieved passages are sent to the configured model provider.

See the User guide for complete instructions, backups, and fully local model configuration.

Repository layout

SciFlow/
├── src/                  # React frontend
├── app/                  # FastAPI backend
├── alembic/              # Database migrations
├── tests/                # Automated tests
├── docs/zh-CN/           # Chinese user and developer guides
├── docs/en/              # English user and developer guides
├── Dockerfile            # Multi-stage frontend/backend image
├── compose.yaml          # Complete local application
├── compose.dev.yaml      # Development database port override
├── start-local-*         # Bilingual cross-platform startup entry points
└── stop-local-*          # Bilingual cross-platform stop entry points