Skip to main content
Nimble Search reads the live web in real time and returns structured, agent-ready results. Every answer reflects what’s on the page right now, not a snapshot from earlier. Pick how much content comes back with search_depth.

Perform a Search with Nimble

Get an API key from the Nimble Platform and install a client:
1

Make a search call

2

Read the response

No search_depth was set, so this used the default, standard. Each result carries a title, URL, description, and metadata:

Search Depth

The search_depth parameter controls how much content each search returns. Choose the right depth for your use case. Need the full page? Set full_content: true alongside either depth. It adds each result’s full scraped page content on top of its title, URL, and description, at higher cost.

Search Depth Best Practices

Learn when to use each depth mode, optimization tips, and cost strategies.

Parameters

API Parameters

For detailed parameter documentation, see the Search API Reference.

Key Features

Add full_content: true to return each result’s complete scraped page content. Works with either search_depth value, at a higher cost.
Filter results by recency using time_range or specify exact date ranges. Essential for news monitoring, tracking recent developments, or analyzing historical data within specific timeframes.
Filter results to specific domains or exclude unwanted ones. Focus on trusted sources, exclude noise (social media, ads), or search within curated lists of authoritative sites.
Need the same domain rules on every request? Nimble can set them for your account or a single API key. See Guardrails.
Account governance rules can override this. See Governance.
Search for specific document types like PDFs, spreadsheets, or presentations (only works with focus: "general"). Perfect for academic research, finding reports, whitepapers, or structured data files.
content_type only works with focus: "general". Supported formats: pdf, docx, xlsx, pptx, and semantic groups: documents, spreadsheets, presentations.
Filter out toxic content from results by setting safe_search. Perfect for customer-facing agents where input can be unexpected.
cURL
This parameter is optional. To apply a level to every request without sending it, see Guardrails.Account governance rules can override this. See Governance.Response behaviorFiltering can remove every match, so a successful 200 response can contain an empty results array. The response’s safe_search field returns the effective level that was applied:
If filtering is temporarily unavailable, Search returns 502 instead of unfiltered results. Retry the request.

Focus Modes

Focus modes narrow retrieval to the content type your agent or application needs and can return structured entities about the selected vertical when available.

Available Modes

Set focus to one mode matching your use case.
general, company and creators support both lite and standard search depth. Other focus modes require search_depth="lite".
Find companies matching a market and geography:

Best Practices

Focus Modes best practices

Custom Focus Mode

For advanced use cases, you can explicitly specify which Extract Templates to use by passing an array of template names to the focus parameter. This gives you full control over data sources and enables mixing templates from different focus modes. Example:

Available Extract Templates

View the complete list of Extract Templates you can use in custom focus modes.

Search vs other tools

Use cases

Research & Data Collection

Gather comprehensive information on any topic from multiple sources automatically

AI Agent Tasks

Get quick, structured web results for your applications

Content Monitoring

Track mentions, news, or updates about specific topics, brands, or keywords

Competitive Intelligence

Monitor what’s being said about competitors or track industry trends

Next steps

API Reference

Explore endpoints, request parameters, and response schemas

Web Tools Skill

Add real-time web intelligence tools to Claude Code and other AI agents

LangChain Integration

Connect Nimble Search to your LangChain applications