Perplexity AI

Perplexity AI is an AI-powered search and answer engine founded in 2022 by Aravind Srinivas, a former research scientist at OpenAI. It was created to bridge the gap between traditional search engines and conversational AI chatbots. Unlike standard chatbots, Perplexity is designed to deliver fast, accurate, and well-sourced information in response to natural language questions by performing real-time web searches and citing its sources.

The platform quickly gained traction for its ability to provide up-to-date information, a significant advantage over models like ChatGPT 3.5, which was initially limited to data up to September. Key milestones include the introduction of Perplexity Copilot for guided research, Perplexity Pro with access to multiple advanced models (GPT-4, Claude, etc.), and the launch of Spaces for team collaboration. By 2025, it had expanded its capabilities to include image generation, file uploads for analysis, and an AI voice assistant for iOS.

Features and Functionality

Perplexity AI combines the intuitiveness of a chatbot with the power of a research engine:

• Real-Time Web Search: Connects to the open internet to pull the most current information from sources like news sites, academic journals, and Reddit, providing answers with citations and links.

• Multi-Model Access (Pro Feature): Subscribers can choose from state-of-the-art large language models (LLMs) including OpenAI’s GPT-4, Anthropic’s Claude 3, and others to power their queries, tailoring the response style and capability to their needs.

• Perplexity Copilot: An interactive guide that helps refine broad research topics through follow-up questions, making it ideal for deep dives into complex subjects.

• File Upload & Analysis: Users can upload PDFs, CSV files, and text documents (e.g., research papers, datasets) and ask Perplexity to summarise or analyse the content within them.

• Focus & Source Selection: Allows users to narrow searches to specific types of sources, such as academic journals, web results, or even a particular site like PubMed, ensuring higher relevance and credibility.

• Spaces for Collaboration: A feature for teams to organise research, share findings, and collaborate on projects within a shared workspace.

• AI Image Generation (Pro Feature): Integrated access to models like DALL-E and Stable Diffusion XL allows Pro users to generate images directly within the platform.

• iOS Voice Assistant: Enables voice-activated queries and tasks, such as sending emails or making reservations, directly from Apple devices.

Pros & Cons Table

Overall Rating

4.2 / 5 ★★★★☆

Perplexity AI excels as a research and information retrieval tool, setting itself apart with its citation of sources and real-time web access. It is arguably superior to traditional chatbots for factual queries and deep research. However, it is not immune to AI hallucinations, and its premium features come at a cost. User experience is also marred by significant customer support and billing issues reported by a portion of its user base.

Versus Competitors:

• vs. Google Search: Provides summarised, cited answers instead of a list of links, saving time, but can be less comprehensive for highly niche queries.

• vs. ChatGPT (free): More useful for research requiring current information and verifiable sources, but ChatGPT may be more creative for open-ended conversation and content generation.

Key Reviews with Links

1. CNET (2025) – Mixed

   • Verdict: “Imagine ChatGPT with an Internet Connection.” Praises its ability to pull from Reddit and link to sources but notes it can falter at synthesising information for difficult queries and can hallucinate.

   • Read full review

2. Trustpilot (1.5/5 | Mostly Negative) – 2025

   • Verdict: The overall rating is low. Users report “terrible customer service”, unexpected charges, refund refusals, and account cancellations without clear explanation. Some also note buggy performance and slow response times.

   • See user reviews

3. Team-GPT (In-Depth Review) – 2025

   • Verdict: 7.7/10. Praises the range of models, intuitive UI, and research capabilities but criticises the image privacy loophole and limited team collaboration features, giving security a 5/10 rating.

   • Read full review

4. University of Tennessee LibGuides – Academic Perspective

   • Verdict: Highlights the “Academic Focus” feature for searching scholarly sources and the value of “Deep Research” but cautions that it uses relatively few sources and may hallucinate or misrepresent citations.

   • Read overview

Summary: Key Points

• 🏆 Core Strength: Sourcing and citation. Its greatest value is providing verifiable, up-to-date answers with links to origins, making it a powerful research assistant.

• 🚀 Key Differentiator: Real-time web search with a conversational interface. It feels like chatting with a knowledgeable expert who can browse the web for you.

• 🎯 Ideal For:

  • Students, academics, and professionals conducting in-depth, source-based research.

  • Anyone needing quick, summarised answers to current events or complex questions without sifting through search engine pages.

  • Casual users looking for a free, ad-free alternative to traditional search for general knowledge.

• ⚠️ Important Considerations:

  • Never treat its output as absolute truth. Always verify the provided sources, as it can hallucinate or misinterpret citations.

  • Be extremely cautious about uploading any sensitive or private files/images due to the identified security risk.

  • Be aware of the mixed user reviews regarding customer support and billing practices before subscribing to a paid plan.

• 💰 Pricing Transparency:

  • Free Plan: Very capable, with 5 Pro searches every 4 hours and unlimited quick searches.

  • Pro Plan: $20/month or $200/year for unlimited Pro searches, multi-model access, image generation, and unlimited file uploads10.

  • Enterprise Pro: Custom pricing for large teams with enhanced security and admin features.

• 🛠️ Ecosystem: Offers mobile apps, browser extensions, and API access (with credits on the Pro plan) for integration into other workflows.