Perplexity基于深度学习的对话式 AI 搜索引擎，提供实时来源的答案。

4.4 (5)

审阅者 Daniel Nikulshyn·更新 2026年5月

AI Search Research Conversational AI Citations Web Search Productivity Freemium

概览

Perplexity 是一款 AI 驱动的答案引擎，结合大型语言模型与实时网页搜索来回答自然语言问题。它不会仅返回一系列链接，而是给出直接答案并附上行内引用，便于快速核实信息来源。用户可以在分线程对话中提问跟进问题，探索相关查询，并切换不同的底层模型。Perplexity 可通过网页、移动应用和浏览器扩展使用，提供免费层和 Pro 计划，后者解锁更高级的模型、文件上传及更深入的研究功能。

主要功能

实时 WEB 搜索与 AI 选摘
答案每篇都有inline引用
Pro 搜索多阶段研究
选择后端模型（GPT，Claude 等）
文件和文档上传进行分析
浏览器扩展和移动应用

价格

模型: Free
分类: Research
评分: 4.4 / 5 (5)

使用场景

事实检查迅速调查

通过 inline 引用来询问自然语言问题并获得答案，轻易验证声明并追踪信息来源于原始 WEB 资源.

多阶段主题探索

在 Pro 搜索和后续问答线索中深入主题，通过询问相关角度的线索单次进行话语.

文档分析与上下文 WEB

在 Pro 版上载上传文件并使用Perplexity总结、分析或回答内容，结合文档内容与实时 WEB 信息

浏览器或移动上进行即时答复

用浏览器扩展或移动应用快速寻找引用的回答同时阅读文章、浏览 WEB 或工作在远程场景下.

优点 & 缺点

优点

答案提供来源引用
实时从 WEB 中拉取信息
清晰的对话式界面
免费版功能足够
支持后续跟进和线索式查询
支持多种后端 LLM

缺点

答案可能仍然存在不准确
来源质量依topic 变化
高级模型需要付费
难以定制化比自己搭建 RAG pipeline

评测

4.4

5 个评分的平均值。

登录以留下评测。

Olga Ivanova

May 10, 2026

Compared a few options

Evaluated this against two competitors. Where it wins: inline citations for every answer and pulls in real-time information from the web. On balance the feature set — especially inline citations for every answer — justifies the 5 stars for our use case.

Liam O’Connor

May 4, 2026

Years in this space

I've evaluated a lot of these over the years. What stands out here is choice of underlying LLMs (GPT, Claude, etc.) — handled better than most — and free tier is genuinely useful. Worth the time if this is your use case.

Ingrid Bauer

Feb 20, 2026

Solid for our team

We rolled this out across the team last quarter and pulls in real-time information from the web. Real-time web search with AI summaries fits neatly into how we already work, and browser extension and mobile apps removed a step we used to do by hand. Less customizable than building your own RAG pipeline, which is the main caveat, but it has held up under daily use.

Linda Petersen

Sep 17, 2025

Solid for our team

We rolled this out across the team last quarter and supports follow-up and threaded queries. Choice of underlying LLMs (GPT, Claude, etc.) fits neatly into how we already work, and choice of underlying LLMs (GPT, Claude, etc.) removed a step we used to do by hand. Less customizable than building your own RAG pipeline, which is the main caveat, but it has held up under daily use.

Mei-Ling Wong

Jun 16, 2025

Does the job

Pretty happy overall. Real-time web search with AI summaries just works and supports follow-up and threaded queries. Less customizable than building your own RAG pipeline can be annoying, but no dealbreakers — I'd recommend it to a friend without hesitating.