crawldown

Crawl any website and save every page as organized Markdown files.

crawldown mirrors a website's URL structure into a local directory of .md files — perfect for archiving documentation, feeding content into RAG pipelines, or reading offline.

crawldown https://docs.example.com --output ./docs-mirror

docs-mirror/
├── index.md
├── getting-started/
│   ├── index.md
│   └── installation.md
└── api/
    ├── reference.md
    └── authentication.md

Installation

pip install crawldown

Or with uv:

uv tool install crawldown

First-time setup — crawldown uses crawl4ai under the hood, which needs a browser installed once to handle JavaScript-rendered pages:

crawl4ai-setup

Quickstart

CLI

# Crawl an entire site
crawldown https://docs.example.com --output ./output

# Limit crawl depth
crawldown https://docs.example.com --output ./output --depth 2

# Add a delay between requests (seconds)
crawldown https://docs.example.com --output ./output --delay 0.5

# Only crawl URLs matching a pattern
crawldown https://docs.example.com --output ./output --include '/docs/*'

# Skip URLs matching a pattern
crawldown https://docs.example.com --output ./output --exclude '/api/*'

# Skip robots.txt enforcement
crawldown https://docs.example.com --output ./output --no-robots

Python API

import asyncio
from crawldown import crawl

asyncio.run(crawl("https://docs.example.com", output_dir="./output"))

With options:

import asyncio
from crawldown import crawl
from crawldown.models import CrawlConfig

config = CrawlConfig(
    url="https://docs.example.com",
    output_dir="./output",
    max_depth=3,
    delay=0.5,
    respect_robots=True,
    include=["/docs/*"],   # only crawl these paths (glob)
    exclude=["/api/*"],    # skip these paths (glob)
)

asyncio.run(crawl(config))

To crawl specific subpages directly, use max_depth=0. Output paths are always anchored to the site root, so files land in the right place regardless of where the crawl starts:

config = CrawlConfig(
    url="https://example.com/privacy-policy",
    output_dir="./output",
    max_depth=0,
    # Writes to ./output/privacy-policy/index.md  ✓
)

How it works

1. Starts at the given URL and fetches the page using crawl4ai (handles JavaScript-rendered pages).
2. Extracts all links that stay within the same domain.
3. Converts each page to Markdown and saves it at a path matching the URL structure.
4. Repeats for every discovered link up to max_depth (default: unlimited).

Options

Contributing

We welcome contributions of all kinds. See CONTRIBUTING.md for how to get started.

License

MIT — see LICENSE.