What is agentic search?
Agentic search uses an AI agent to plan and execute multiple coordinated search queries rather than submitting a single static query. The agent evaluates what each result set contains, identifies gaps, reformulates queries to fill those gaps, and repeats until it has collected sufficient sources to answer the original question. Unlike a single web search API call that returns results once, agentic search is a loop: each round of results informs the next query.
| Factor | Standard search | Agentic search |
|---|---|---|
| Query count | One | Multiple, adaptive |
| Query reformulation | Manual | Automatic based on prior results |
| Source coverage | What one query surfaces | Cross-checked across many angles |
| Latency | Low | Higher, scales with depth |
| Best for | Lookup and discovery | Complex questions, research synthesis |
Use agentic search for questions that no single query can answer fully: competitive intelligence requiring synthesis across dozens of pages, due diligence involving cross-checking claims across sources, or research where terminology varies by community. For straightforward lookups where the first result set is sufficient, a single search API call is faster and cheaper.
Firecrawl's /agent endpoint handles agentic search end-to-end: it plans queries, navigates to sources, extracts content, and returns structured results without requiring orchestration code on the caller's side. For a deeper dive into agentic search architecture — including RAG integration, real-time retrieval patterns, vector database usage, and implementation examples — see the dedicated guide.