This OpenClaw skill integrates Decodo's Web Scraping API into any OpenClaw-compatible AI agent or LLM pipeline. It exposes two tools that agents can call directly:
google_search– query Google Search and receive structured JSON resultsuniversal– fetch and parse any public webpage, returning clean Markdown
Backed by Decodo's residential and datacenter proxy infrastructure, the skill handles JavaScript rendering, bot detection bypass, and geo-targeting out of the box.
- Real-time Google Search results scraping
- Universal URL scraping
- Structured JSON or Markdown results
- Simple CLI interface compatible with any OpenClaw agent runtime
- Minimal dependencies — just Python with Requests
- Authentication via a single Base64 token from the Decodo dashboard
- Python 3.9 or higher
- Decodo account with access to the Web Scraping API
- OpenClaw installed on your machine
- Clone this repo.
git clone https://github.com/Decodo/decodo-openclaw-skill.git
- Install dependencies.
pip install -r requirements.txt
- Set your Decodo auth token as an environment variable (or create a
.envfile in the project root):
# Terminal
export DECODO_AUTH_TOKEN="your_base64_token"
# .env file
DECODO_AUTH_TOKEN=your_base64_token
This skill ships with a SKILL.md file that defines both tools in the OpenClaw skill format. OpenClaw-compatible agents automatically discover and invoke the tools from this file without additional configuration.
To register the skill with your OpenClaw agent, point it at the repo root — the agent will read SKILL.md and expose google_search and universal as callable tools.
Search Google and receive a structured array of results:
python tools/scrape.py --target google_search --query "your query"
Fetch and convert any webpage to clean Markdown file:
python tools/scrape.py --target universal --url "https://example.com/article"
Decodo Web Scraping API documentation
ClaWHub – OpenClaw skill registry
All code is released under the MIT License.