What is agentic search?

Agentic search is search driven by an AI agent that iterates: it searches, evaluates results, refines queries, and repeats until it finds the right sources.

Instead of a single query, an agent plans multiple search steps and adapts based on what it finds. This is common in deep research and AI agent workflows that use a Web Search API for discovery.

Use agentic search when the first results are incomplete, when the topic is broad, or when you need to cross-check sources. It is useful for building source lists, finding long-tail pages, and filling gaps before extraction.

Agentic search turns search into a multi-step, goal-driven loop. Firecrawl's /agent endpoint searches, navigates, and extracts in one flow when you need end-to-end automation.