rawfy 0.1.5

Python wrapper for the Rawfy AI agent skill — converts any URL into structured, agent-readable content

Rawfy Python Wrapper

Python bindings for the Rawfy AI agent skill.

Prerequisites

• Node.js >= 18 — Rawfy is a Node.js tool
• Rawfy CLI — npm install -g rawfy

Install

pip install .

Quick Start

from rawfy import fetch, metadata

# Full page fetch (returns WSM markdown)
content = fetch("https://example.com")

# Get structured JSON data
import json
data = fetch("https://example.com", format="json")
parsed = json.loads(data)

# Or use the convenience wrapper
from rawfy import fetch_json
data = fetch_json("https://example.com")
print(data["metadata"]["title"])

# Lightweight metadata only
meta = metadata("https://docs.python.org")
print(f"Title: {meta['title']}, Words: {meta['word_count']}")

API

fetch(url, *, format="markdown", vision=False, no_playwright=False, max_tokens=50000, timeout=30)

Fetch and process a URL. Returns the content as a string.

fetch_json(url, **kwargs)

Same as fetch() with format="json", returns a parsed dict.

metadata(url, *, no_playwright=False, timeout=15)

Fetch only page metadata (lightweight). Returns a dict.

check_installation()

Check if Rawfy and dependencies are installed. Returns a status dict.

Error Handling

from rawfy import fetch, RawfyError

try:
    content = fetch("https://example.com")
except RawfyError as e:
    print(f"Error [{e.code}]: {e}")
    print(f"URL: {e.url}")

CLI

python -m rawfy https://example.com
python -m rawfy https://example.com --format json