Tutorial: First Crawl
Goal​
Run AutoPOM-Agent on a public site and inspect generated outputs.
Steps​
PYTHONPATH=src python3 -m autopom.cli.main \
--base-url "https://example.com" \
--output-dir "output" \
--pom-language "javascript" \
--max-depth 2 \
--max-pages 20
Validate outputs​
- Open
output/models_jsonfor extracted page models. - Open
output/javascript/pagesfor generated classes. - Open
output/reports/crawl_summary.mdfor confidence metrics.
Try Interactive Mode​
For more control over what is captured:
PYTHONPATH=src python3 -m autopom.cli.main \
--interactive \
--pom-language "typescript"
This allows you to verify the page state exactly as the agent sees it before it begins mapping.
Repeat with TypeScript​
PYTHONPATH=src python3 -m autopom.cli.main \
--base-url "https://example.com" \
--output-dir "output" \
--pom-language "typescript" \
--max-depth 2 \
--max-pages 20