AI APP PROFILE

Perplexity

The fastest way from a question to the sources behind the answer, with the weakest published privacy position of the group.

Source review Last reviewed August 7, 2026 By Perplexity AI

The facts at a glance

ProviderPerplexity AI
TypeAI search and research
Free optionYes, with limited weekly usage and basic models only
Paid pricePro USD 20 per month, Max USD 200 per month, both billed monthly
AvailabilityWeb, iOS, Android, browser extension

Pricing last checked August 7, 2026. Prices change and vary by region. Check the official pricing page ↗

Assessment

What it does well

  • Answers are built around citations rather than having them bolted on
  • Very fast at surfacing current web material
  • Low friction follow up questions make it good for exploring a topic rather than just querying it

Where it falls short

  • Citations sometimes support only part of a generated sentence, so you still have to read them
  • The free tier restricts you to basic models with limited weekly usage
  • Perplexity's privacy notice says data is used to improve or create services and products including its AI models, and we found no dedicated training opt out for general queries. The only explicit carve out is that email content is not used for training

Privacy and data control

Perplexity states data is used to improve or create its services and products, including its AI models. We found no dedicated opt out toggle for general queries. The one explicit exclusion is that email content is not used to train models.

Read the vendor privacy page ↗

The longer read

The 18BYTE view

Perplexity makes the research trail visible. Its interface encourages users to inspect sources, refine a question, and move quickly from a summary into the material behind it.

Choose Perplexity if

Your work begins with the open web and your first question is usually, “What do the best available sources say?”

Look elsewhere if

You need a single environment for extensive drafting, creative production, coding, and longer multi-stage projects.

What to watch

Known limitations and defaults worth knowing before you commit. Each one says where it came from and when we last checked it.

  1. A citation does not always support the whole sentence

    Editorial watchout
    What can happen
    Citations sometimes support only part of the sentence they are attached to, so the presence of a source is not the same as the claim being sourced.
    Why it matters
    The entire value of a citation-first tool is that you can trust the link. Where the link covers half the sentence, the interface still looks fully sourced.
    Who should be careful
    Anyone quoting the output, or using it as the evidence base for something published or filed.
    What to do about it
    Open the citation and check it covers the specific claim, not the general topic. Treat the citation as a pointer to read, not as verification already done.
    18BYTE editorial judgment, no vendor documentation to cite Last verified August 7, 2026

  2. We found no training opt out for general queries

    Documented warning
    What can happen
    Perplexity states data is used to improve or create its services and products, including its AI models. We found no dedicated opt-out toggle for general queries. The one explicit exclusion is that email content is not used to train models.
    Why it matters
    On the other assistants there is at least a switch. Here, on what we could verify, there is not one for ordinary queries, so the only control available is what you choose to type.
    Who should be careful
    Anyone researching a confidential subject, where the query itself is the sensitive part.
    What to do about it
    Assume the query is retained. Keep identifying context out of the search text, and check the privacy policy for changes before relying on it for confidential research.

How it performs by outcome

We assess products against a job, not a leaderboard. These are the outcomes this entry is compared under.

Worth comparing against

Grok Questions about what is happening right now on the open web and on X Microsoft Copilot Spreadsheet and document work for people already paying for Microsoft 365 Notion AI Answering questions from a team wiki rather than from the open web Claude Fable 5 Long horizon tasks where the cost of a wrong answer exceeds the cost of the tokens

How this profile was produced

Source review. Every fact here is drawn from the provider documentation linked below and checked on the date shown. We have not yet run this entry through a controlled test, and no scoring is implied.

Read the full review methodology →

Primary sources