Advanced Web Search
What can Advanced Web Search do?
Advanced Web Search gives you powerful research capabilities with fine-grained control over how searches are performed. Unlike Standard Web Search, it can scrape full web pages, analyze content using AI, and provide in-depth summaries.
Best for:
- In-depth research projects
- Analyzing content from multiple sources
- Finding detailed information within web pages
- Research that requires understanding full context, not just snippets
Key differences from Standard:
- Scrapes complete web pages (optional)
- Uses AI to summarize content (optional)
- Searches within scraped pages for relevant sections
- Returns structured, comprehensive research results
How to enable Advanced Web Search
Using the Playground
Advanced Web Search is called "Deep Research" in the Playground interface.
- Navigate to the Playground page
- Select your AI model
- Look for the rocket icon (🚀) with "Deep Research" label
- Click the Deep Research button to enable it
- Active state: Button shows darker color and border
- Inactive state: Button shows lighter color
- Type your research question
- Click Send
Note: You can enable both Web Search (globe icon) and Deep Research (rocket icon) at the same time for maximum research capability.
Using the Teammate Builder
For Teammates that need advanced research capabilities:
- Open the Teammate Builder
- Find "Web Search (Advanced)" in the Tools section
- Drag it onto the canvas
- Connect it to your Teammate node
- Configure the advanced options (see below)
- Save your Teammate
Configuring Advanced Web Search
Advanced Web Search has more options to customize your research process.
Basic Settings
Limit
What it does: Number of Google search results to retrieve
- Range: 10 to 100 results
- Default: 10 results
- Recommendation: Use 10-20 for most research tasks
Additional Search Terms (Optional)
What it does: Same as Standard Web Search - refines your search
Examples:
site:edu- Only search educational institutionsfiletype:pdf- Find PDF documents2024..2025- Search within a date range
Web Scraping Settings
These options appear when you enable page scraping.
Scrape Result Pages?
What it does: Downloads and analyzes the full content of search result pages
- Default: Off (disabled)
- When to enable: When you need more than just search result snippets
- Impact: Makes searches slower but provides deeper information
Maximum Scraped Pages
What it does: Limits how many pages to scrape (only visible when scraping is enabled)
- Range: 1 to 100 pages
- Default: 5 pages
- Recommendation: Use 5-10 pages for balance between depth and speed
- Important: More pages = slower search, but more comprehensive results
Matches Per Scraped Page
What it does: How many relevant sections to extract from each scraped page
- Range: 1 to 20 sections
- Default: 3 sections
- Recommendation: Start with 3, increase if you need more detail
- Impact: Higher numbers provide more context but slower processing
Summarization Settings
These options appear when both scraping and summarization are enabled.
Summarize Scraped Pages
What it does: Uses AI to create concise summaries of scraped content
- Default: On (enabled) when scraping is enabled
- Benefit: Reduces information overload and focuses on relevant details
- Recommendation: Keep this enabled for most research tasks
Model Name
What it does: Selects which AI model to use for summarization (required when summarization is enabled)
- Options: Dropdown shows all available AI models
- Recommendation: Choose a model that balances speed and quality for your needs
Temperature
What it does: Controls how creative vs. focused the summaries are
- Range: 0 to 1
- Default: 0.3
- Lower values (0-0.3): More focused, factual summaries
- Higher values (0.7-1.0): More creative, varied summaries
- Recommendation: Keep at 0.3 for research tasks requiring accuracy
Configuration examples for different use cases
Quick research (fast, less detailed)
✓ Scrape Result Pages: ON
- Maximum Scraped Pages: 3
- Matches Per Scraped Page: 2
✓ Summarize Scraped Pages: ON
- Temperature: 0.3
Comprehensive research (slower, very detailed)
✓ Scrape Result Pages: ON
- Maximum Scraped Pages: 10
- Matches Per Scraped Page: 5
✓ Summarize Scraped Pages: ON
- Temperature: 0.2
Fast facts only (no scraping)
✗ Scrape Result Pages: OFF
(All other options hidden)
What you'll get back
When Advanced Web Search runs with full features enabled, you'll receive:
- Search results with titles and links
- Original snippets from Google
- Extracted sections from scraped pages (if scraping enabled)
- AI-generated summaries of relevant content (if summarization enabled)
- Source citations for all information
The Teammate uses this comprehensive research to provide thorough, well-sourced answers.
Troubleshooting Advanced Web Search
Problem: Search is very slow
- Cause: Scraping too many pages or extracting too many matches
- Solution: Reduce "Maximum Scraped Pages" to 5 or fewer, or reduce "Matches Per Scraped Page" to 2-3
Problem: Getting too much information
- Cause: Too many matches per page or too many pages being scraped
- Solution: Lower the "Matches Per Scraped Page" setting, or enable "Summarize Scraped Pages" if it's off
Problem: Summaries are too creative or off-topic
- Cause: Temperature setting is too high
- Solution: Lower the Temperature to 0.2-0.3 for more focused summaries
Problem: Not enough detail in results
- Cause: Not enough pages being scraped or not enough matches extracted
- Solution: Increase "Maximum Scraped Pages" or "Matches Per Scraped Page" settings
Problem: Search fails with an error
- Cause: Network timeout or API issue
- Solution: Try again, or reduce the scope of your search (fewer pages, fewer matches)
Updated on: 03/11/2025
Thank you!