Perplexity
Also known as: Perplexity AI, Perplexity search
Perplexity is where many builders default for quick research that needs to be current and cited. Unlike asking a language model a factual question (which risks confident-sounding hallucination), Perplexity searches the live web, reads the results, and synthesizes an answer with inline citations. You can click through to verify any specific claim.
The product has expanded from a simple search alternative to include Deep Research mode (for longer, more thorough multi-step research), Spaces (project-based contexts where you can load specific sources), and an API for developers building research-augmented applications.
Perplexity sits in the same conceptual space as grounding and RAG: the answer is only as good as the sources retrieved. It's most reliable for current factual questions and least reliable for nuanced judgment calls where the web itself may not have good information. For builders, it's both a tool to use directly and a model of how search-augmented AI should behave.