Self-hosted browser scraper that archives liked/starred content from X, Hacker News, Reddit, and GitHub.

• Multi-platform support: X, Hacker News, Reddit (via browser automation), GitHub (via CLI)
• Web dashboard: FastAPI-based UI for managing accounts and browsing archived items
• File mirroring: Items saved as markdown files with YAML frontmatter
• Scheduled scraping: Automatic periodic sync with configurable intervals
• Encrypted credentials: Passwords encrypted with Fernet before storage

• Python 3.11+
• uv package manager
• GitHub CLI (gh) authenticated for GitHub stars

# Clone and install
git clone <repo-url>
cd keep-my-social-data
uv sync

# Configure data directory
cp .env.example .env
# Edit .env to set your preferred DATA_DIR

# Start the dev server
uv run kmsd-dev

# Open http://localhost:8001 in your browser

• GitHub: Run gh auth login first, then add via the web UI
• X/HN/Reddit: Add username and password through the web UI

• Database: SQLite at $DATA_DIR/kmsd.db
• Markdown mirror: $DATA_DIR/{platform}/*.md
• Secrets: $DATA_DIR/secrets.env (auto-generated)

# Run tests
uv run pytest

# Run specific test file
uv run pytest tests/test_github.py

MIT