| name | apify |
| description | Use when scraping websites, extracting structured data from listing sites, running prebuilt scrapers/actors, or automating stubborn web research workflows. Make sure to use this whenever apartment hunting, marketplace scraping, anti-bot-resistant extraction, or batch website data collection would benefit from a dedicated scraping platform rather than ad hoc browser clicks. |
| mcp | {"apify":{"command":"secrets","args":["APIFY_TOKEN","--","npx","-y","@apify/actors-mcp-server"],"env":{"SOPS_AGE_KEY":"${SOPS_AGE_KEY}"}}} |
Apify
Use Apify as the heavy-duty scraping and extraction layer.
It is especially useful when:
- a site is JS-heavy or anti-bot-prone
- we need many listings, not just one page
- a community Actor already exists for the target site
- we want structured output instead of manual copy/paste from the browser
Best use cases
- Apartment / real-estate site scraping
- Airbnb / booking / marketplace extraction
- Bulk listing collection across multiple sites
- Converting messy listing pages into structured datasets
MCP usage
Call via:
skill_mcp(mcp_name="apify", tool_name="...", arguments={...})
Recommended flow
- Search for a relevant Actor.
- Inspect the Actor details and inputs.
- Run the Actor.
- Read the dataset items.
- Normalize results into a shortlist.
Typical tools
Search for a scraper
skill_mcp(mcp_name="apify", tool_name="search-actors", arguments={"query": "airbnb scraper"})
Use the built-in web extraction path
skill_mcp(mcp_name="apify", tool_name="apify/rag-web-browser", arguments={"url": "https://example.com/listing"})
Run a chosen Actor
skill_mcp(mcp_name="apify", tool_name="call-actor", arguments={
"actorId": "apify/example-actor",
"input": {"startUrls": [{"url": "https://example.com"}]}
})
Read results
skill_mcp(mcp_name="apify", tool_name="get-dataset-items", arguments={"datasetId": "DATASET_ID"})
Notes
- Requires
APIFY_TOKEN in the environment.
- Apify is best when we need scale, repeatability, or structured output.
- Prefer Apify over fragile one-off browser scraping when the task looks repeatable.
