Fresh

A CLI tool to fetch the latest documentation from any website in Markdown format.

Requirements

• Python 3.12+
• uv (recommended) or pip

Installation

# Using uv (recommended)
uv sync

# Or using pip
pip install -e .

Quick Start

# List all documentation pages on a website
fresh list https://docs.python.org/3/

# Fetch a documentation page and convert to Markdown
fresh get https://docs.python.org/3/tutorial/

# Search for content across documentation pages
fresh search "virtual environment" https://docs.python.org/3/

# Download entire documentation for offline use
fresh sync https://docs.python.org/3/

# Add an alias for quick access
fresh alias add python https://docs.python.org/3/
fresh list python

Commands

fresh list

List all documentation pages available on a website.

fresh list <URL> [OPTIONS]

Options:

• -v, --verbose - Use rich output format
• -p, --pattern TEXT - Filter paths matching pattern
• -d, --depth INTEGER - Maximum crawl depth (default: 3)
• --max-pages INTEGER - Maximum number of pages to discover (default: 100)
• --sort TEXT - Sort results by name or path
• -f, --format TEXT - Output format: json, yaml, xml
• -c, --count - Show only total count

fresh get

Fetch a documentation page and convert it to Markdown.

fresh get <URL> [OPTIONS]

Options:

• -v, --verbose - Use verbose output
• -t, --timeout INTEGER - Request timeout in seconds (default: 30)
• --header TEXT - Custom HTTP header
• --no-follow - Do not follow redirects
• --skip-scripts - Exclude JavaScript from output
• --no-cache - Bypass cache
• -o, --output TEXT - Write output to file
• -r, --retry INTEGER - Number of retry attempts (default: 3)
• --dry-run - Show what would be fetched without downloading

fresh search

Search for content across documentation pages.

fresh search <QUERY> <URL> [OPTIONS]

fresh websearch

Search the general web for any topic.

# Basic search (uses DuckDuckGo HTML, free)
fresh websearch "python async tutorial"

# With specific number of results
fresh websearch "react hooks" --count 5

# Table output for human-readable display
fresh websearch "rust ownership" --table

Options:

• -n, --count INTEGER - Maximum number of results (default: 10)
• -e, --engine TEXT - Search engine: auto, ddg, brave (default: auto)
• -j, --json - Output as JSON (default)
• -t, --table - Output as table
• -v, --verbose - Show verbose output

Note: Brave Search API can be used by setting the BRAVE_API_KEY environment variable. DuckDuckGo is used by default (free, no API key required).

fresh alias

Manage library aliases for quick access.

# Add an alias
fresh alias add <name> <url>

# List all aliases
fresh alias list

# Search aliases
fresh alias search <query>

fresh sync

Download entire documentation for offline use.

fresh sync <URL> [OPTIONS]

Options:

• -o, --output-dir PATH - Target directory for synced docs
• -v, --verbose - Use verbose output
• --max-pages INTEGER - Maximum number of pages to sync (default: 100)
• -d, --depth INTEGER - Maximum crawl depth (default: 3)
• -f, --force - Force re-sync (delete existing files first)
• -p, --pattern TEXT - Filter paths matching pattern

fresh guide

Manage personal guides and documentation.

# Create a new guide
fresh guide create my-guide --content "Guide content here" --title "My Guide"

# List all guides
fresh guide list

# Show a specific guide
fresh guide show my-guide

# Search across guides
fresh guide search "keyword"

Configuration

Fresh stores its configuration in the following locations:

• Aliases: ~/.fresh/aliases.json
• Cache: ~/.fresh/cache/
• Sync data: ~/.fresh/docs/
• History: ~/.fresh/history.db

Development

# Install dev dependencies
uv sync --extra dev

# Run linter
ruff check .

# Run type checker
mypy src/

# Run tests
pytest

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Author

• Nesalia Inc.

Security

If you discover any security vulnerabilities, please send an e-mail to support@nesalia.com.

License

MIT License - see the LICENSE file for details.