Deploy and Host AutoFigure-Edit on Railway

About Hosting AutoFigure-Edit on Railway

AutoFigure-Edit runs as a single Dockerized FastAPI web application on Railway. Railway gives managed deployment, HTTPS domain, and variable management with no server operations.

Tech Stack

• Python 3.11
• FastAPI + Uvicorn
• Pillow and scientific Python libs
• Docker Hub image on Railway

Why Deploy AutoFigure-Edit on Railway

• Single-service deployment with built-in HTTPS
• Fast rollout from a Docker image
• Easy variable-based configuration and redeploy
• Good fit for research demos and internal sharing

Common Use Cases

• Internal scientific figure generation web service
• Team-shared AutoFigure-Edit playground
• Reproducible cloud demo for paper methods

Deployment Notes

• HTTP container port is 8000
• Railway routing is controlled by PORT=8000
• Public domain is enabled for browser/API access
• HF_TOKEN / HUGGINGFACE_HUB_TOKEN should be set for RMBG-2.0 gated model access
• ROBOFLOW_API_KEY and FAL_KEY are optional unless those SAM backends are used

Dependencies for AutoFigure-Edit on Railway

This deployment only requires one application service and no external database.

Deployment Dependencies