AI Search Engine Exa Launches Monitors to Periodically Crawl the Entire Web for Fresh Content

According to 1M AI News, the AI search engine Exa has launched the Monitors feature, allowing developers to set up scheduled search tasks. The search results are automatically deduplicated and pushed via Webhook, returning only new content since the last run. Exa's search API is used by companies such as Cursor, Notion, AWS, and Databricks, positioning itself as a search infrastructure for AI Agents.

Monitors support custom search frequencies (minimum of 1 hour) and allow specifying a structured JSON output format through the outputSchema parameter, enabling automatic extraction of fields like company names, funding rounds, amounts, and the option for a plain text summary. In addition to scheduled triggers, it supports manual ad-hoc runs and provides three access methods: Python, JavaScript, and cURL. Typical use cases include tracking competitor activities, monitoring funding announcements in specific sectors, and following academic paper publications.

Prior to Monitors, AI Agents relied on actively calling the search API for external information. Monitors transform this mode from on-demand queries to subscription-based pushes, allowing Agents to continuously receive incremental information without the need for polling.