novada_scrape — Platform Scraper Skill

novada_scrape — Platform Scraper Skill

When to use: You need structured records from a known platform (Amazon, TikTok, Reddit, LinkedIn, etc.) — not raw HTML, but clean tabular data.

Step 1: Find the right platform and operation

ALWAYS read the novada://scraper-platforms resource before calling novada_scrape. Platform names and operation IDs are exact — guessing causes 11008 errors.

// Read the resource first:
// novada://scraper-platforms

Common platforms quick reference

Platform	Operation example	Key params
amazon.com	amazon_product_by-keywords	keyword, num
amazon.com	amazon_product_by-asin	asin
reddit.com	reddit_posts_by-keywords	keyword, num
tiktok.com	tiktok_user_videos	username, num
linkedin.com	linkedin_job_listings	keyword, location
google.com	google_search	q, num
glassdoor.com	glassdoor_jobs_by-keywords	keyword, location
zillow.com	zillow_listings	location
github.com	github_repository_details	owner, repo

Call pattern

{
  "platform": "amazon.com",
  "operation": "amazon_product_by-keywords",
  "params": {"keyword": "mechanical keyboard", "num": 10},
  "format": "markdown",
  "limit": 10
}

Format guide

• markdown — best for agents reading and reasoning over results
• json — best for code processing and downstream pipelines
• toon — token-optimized format (40-65% smaller), pipe-separated rows

Error handling

Error	Meaning	Action
11006	Scraper API not activated	Activate at dashboard.novada.com. Do NOT retry.
11008	Invalid platform name	Check platform name. Read novada://scraper-platforms.
10001	Invalid file type for operation	Try different operation. Contact support.
code 0, no task_id	Unexpected response shape	Retry once.

When NOT to use novada_scrape

• Arbitrary web pages → use novada_extract
• Pages not in the 129-platform list → use novada_crawl
• After getting 11006 → this is a plan-tier error, don't retry