Exa Core Workflow B

Exa Core Workflow B

Overview

Secondary workflow for Exa. Complements the primary search workflow by focusing on advanced retrieval patterns: similarity search from a seed URL, domain-filtered queries, and batching multiple queries efficiently. Use this skill when you need to find pages similar to a known reference, restrict results to specific domains, or run bulk research queries within your API quota.

Prerequisites

• Completed exa-install-auth setup
• Familiarity with exa-core-workflow-a
• Valid API credentials configured

Instructions

Step 1: Setup

Determine whether to use URL-based similarity search or a domain-scoped query. For similarity search, provide a seed URL that represents the type of content you want to surface. For domain-scoped queries, specify the include_domains or exclude_domains filters to narrow results to authoritative sources.

// Step 1 implementation

Step 2: Process

Execute the configured query using the appropriate Exa endpoint variant. Review the returned results for coverage and quality. If using similarity search, compare the seed URL content against the returned results to confirm the semantic alignment is accurate. Adjust the num_results and start_published_date parameters to refine freshness and volume.

// Step 2 implementation

Step 3: Complete

Aggregate results from multiple queries if running a batch, merge and deduplicate across query outputs, and write the final result set to your target destination. Document the query strategy used so future runs can reproduce or iterate on the approach.

// Step 3 implementation

Output

• Completed Core Workflow B execution
• Filtered or similarity-matched results from Exa API
• Deduplicated, merged result set ready for downstream consumption
• Success confirmation or error details

Error Handling

Examples

Complete Workflow

// Complete workflow example

Error Recovery

// Error handling code

Resources

• Exa Documentation
• Exa API Reference

Next Steps

For common errors, see exa-common-errors.