
by Perplexity
Perplexity is an AI search engine that answers questions in conversational form, using live web search and citing the sources behind each answer.
Web appFreemium
Step 1: Open Perplexity and type a question in plain language.
Step 2: Read the answer, which is written from several web sources.
Step 3: Open the cited links to verify the answer or read the original material.
Step 4: Ask follow-up questions to keep the conversation going.
Step 5: Choose a different AI model, or use Deep Research, when the topic needs more depth.
A traditional search engine gives you many links to sort through. Perplexity writes a direct answer that synthesizes multiple sources and lists them as citations.
Content is sourced from the web in real time as you ask, so answers reflect the latest information available online.
Yes. Perplexity includes models from providers such as OpenAI and Anthropic, and you can select the one that suits your task.
Every response includes citations with links to the original sources, so you can open them and verify the information.
Perplexity Search APIThe Perplexity Search API returns real-time, ranked web results as JSON, with page text extracted into each result. It is for developers who want raw search results to feed their own models, agents or pipelines.
Perplexity ComputerCloud AI agent for research, analysis, file creation, coding, and multi-step work.
Perplexity Hybrid ComputeHybrid AI mode that keeps private-file work on a Mac while using cloud models for research and reasoning.
Motif Research MCPMotif MCP turns a folder of interview transcripts into cited findings and checks them with a critic loop. It runs inside MCP hosts such as Claude Code, Cursor and Claude Desktop.Inquisitor MCPInquisitor is an MCP server and agent skill that makes AI agents solve problems by testing hypotheses instead of retrying blindly. It estimates depth from a cheap first probe and keeps a record of failed ideas.
QY Evolution MCPQianYuan is an identity, trust and result-reuse layer for autonomous agents, reachable through an MCP endpoint. Agents publish results they have already produced, and other agents reuse them to save tokens.