Added

• Public REST API (v1) for AI agent access to scholarly content
• Human-readable scroll URLs at /{year}/{slug} with automatic slug generation at publish time
• Rich social link previews with Open Graph images and structured metadata
• Turnstile CAPTCHA and signup rate limiting on registration
• Per-user upload rate limit of 5 uploads per hour
• Password change feature with forgot-password link
• Account deletion flow
• Infinite draft persistence for academic workflows
• Admin email notifications for new signups and published scrolls
• DOI badge component with Zenodo sandbox detection
• Back-to-top floating action button on scroll view pages
• Post-deployment email health check
• Database health checks with latency metrics
• Backup health check workflow
• Sentry alert rules with abuse-pattern reporting
• Plotly CDN support for interactive widgets

Fixed

• FAB occlusion issues resolved with grouped horizontal layout
• Dashboard hx-boost bug and dead footer partial removed
• Dark mode contrast and How It Works page layout
• Turnstile widget rendering failures with HTMX navigation
• CSRF protection for password change form
• Double navbar after HTMX navigation
• Theme toggle and event listeners for HTMX-loaded content
• Upload form returning full page instead of partial on errors
• Session injection vulnerability via Supabase RLS enforcement
• Email error handling with Sentry integration
• Parallel database conflicts in unit tests
• Flaky mobile login and E2E test timing issues

Changed

• Redesigned floating action buttons as a grouped horizontal unit
• Homepage refactored to HTMX patterns with dark-mode footer fix
• HTML validation switched from regex to BeautifulSoup
• File uploads use UploadFile for better memory efficiency
• Static file cache TTL reduced from 1 year to 1 hour
• Security model relaxed to allow unsafe-eval for interactive research libraries (Bokeh, Plotly)
• Signup form label changed to "Full Name"
• Legal page dates and privacy policy updated with named subprocessors for GDPR compliance
• Excluded html_content from list queries to improve load times
• Unified email verification banners into a reusable component

First public beta release of Scroll Press - a modern, web-native preprint server.

✨ Features

• HTML-native publishing - Upload from Typst, Quarto, MyST, Jupyter, or handwritten HTML
• Preview workflow - Review uploads before publishing
• Authentication - Session-based auth with email verification
• Subject categorization - Organize research by academic disciplines
• Responsive design - Desktop and mobile optimized with dark mode
• GDPR compliance - User data export endpoint
• Performance - Static file caching with CDN-ready headers

🚀 Technical Stack

• FastAPI + PostgreSQL with async SQLAlchemy
• Jinja2 templates with HTMX interactions
• Playwright E2E testing (desktop + mobile)
• Pre-commit hooks for automated linting
• Comprehensive CI with parallel test execution

📦 Deployment

• Production-ready deployment guides
• Fly.io + Supabase infrastructure
• Alembic database migrations
• Resend email integration
• Sentry error tracking
• Automated backups

📚 Documentation

• Complete setup guide (just init)
• Comprehensive testing documentation
• Contribution guidelines
• Deployment instructions

⚠️ Beta Notice

This is a beta release - expect active development and potential breaking changes. We welcome feedback and contributions!

Visit scroll.press to see it in action.