All Docs
FeaturesAgentOS Scope OutUpdated March 12, 2026

Crawl Depth Configuration: Skim vs Full Analysis

Crawl Depth Configuration: Skim vs Full Analysis

When you initiate a crawl from a directory URL, you are prompted to choose a crawl depth. This controls how much work the system does and how long the crawl takes.

Crawl Depth Options

Skim Only

The system traverses the directory, discovers all product listings, and extracts surface-level data (names, categories, brief descriptions, visible metadata). No deep-scraping, scoring, or dossier generation is performed.

  • Duration: Seconds
  • Use when: You want to quickly assess how many listings a directory contains, or whether it's likely to yield useful products, before committing to a full run.
  • Output: A raw listing count and surface data snapshot. No opportunity scores or dossiers are produced.

Full Analysis

The system runs the complete pipeline end-to-end: it skims the directory, then deep-scrapes each product's reviews, ratings, features, and pricing signals, scores each product across all four dimensions (replicability, market demand, revenue potential, competitive gaps), and generates a full dossier for the top-ranked products.

  • Duration: Several minutes (varies by directory size)
  • Use when: You are ready to commit to a full analysis run and want ranked results with dossiers ready to act on.
  • Output: Ranked product dashboard with composite opportunity scores, and per-product dossiers including auto-drafted mission statements, suggested feature lists, and competitor weakness analysis.

Recommended Workflow

  1. Paste a directory URL and select Skim Only.
  2. Review the listing count and surface data to confirm the directory is worth analysing.
  3. Re-run the same URL with Full Analysis to generate the complete scored output.

This two-step approach avoids burning analysis time on directories that turn out to be too small, off-topic, or already well-covered.

Scoring Dimensions (Full Analysis Only)

Full Analysis scores each discovered product across four dimensions:

DimensionWhat It Measures
ReplicabilityHow feasible it is to rebuild with our stack
Market DemandReview volume and sentiment signals
Revenue PotentialPricing tier and addressable market size
Competitive GapsWeaknesses revealed by negative reviews

These four scores are combined into a single composite opportunity score used to rank products in the dashboard.