PerigonInstall

Summary

Perform a semantic vector search for news articles using a natural language prompt and optional filtering criteria.

Inputs

• Prompt (required) — Natural language query string to search articles semantically.
• Additional Options — Optional filters including article IDs, categories, cities, clusters, companies, locations, publication date range, source domains, labels, languages, geographic coordinates with radius, and more, to narrow down search results.
• Advanced Filters — Structured JSON filters with AND, OR, and NOT logic for complex filtering of articles.

Output shape

a list of article objects matching the semantic query and applied filters

Results are delivered in a paginated list up to the specified size (default 10, max 1000). The output includes metadata for each article based on the Perigon API response.

Examples

Example 1: Search for articles related to 'AI technology advancements' filtered by the 'Tech' category and published after 2023-01-01.

Set the Prompt to 'AI technology advancements'; add Additional Options with Category='Tech' and Publication Date From='2023-01-01'; specify desired result count with Size option.