Perplexity

Perplexity is a conversational answer engine founded in San Francisco in August 2022. It retrieves live web results for a question, writes an answer from them, and numbers each sentence back to the page it came from. An academic focus mode narrows retrieval to scholarly sources, and a deep research mode runs an extended multi-search investigation before reporting. It is a general research tool that scientists use rather than a scientific one.
Why it matters
It covers the fast, low-stakes end of research, checking a definition, orienting in an unfamiliar area, chasing down a figure, where opening a bibliographic database is disproportionate.
Live retrieval means it is not bounded by a training cutoff, which matters for anything current.
Inline numbered citations make the sourcing inspectable in a single click, which is the minimum bar for research use.
The academic focus mode narrows the same interface to scholarly sources when the question warrants it.
Key features & advantages
Cited answers over live web retrieval, with the source list shown alongside the response.
Focus modes, including an academic mode restricted to scholarly sources.
Deep Research: an extended mode that runs many searches and returns a longer structured report.
Follow-up questions that carry the retrieval context, so a search becomes a conversation.
Limitations & things to watch
The web is the corpus. Blogs, vendor pages and news sit next to primary literature with no visible difference in standing.
Citations confirm where a sentence came from, not that the source is any good, and misattribution between adjacent sources does happen.
It is not a systematic search, results are not reproducible or exhaustive and should never stand in for a database search in a formal review.
Recency bias in retrieval can bury the foundational paper under this month’s coverage of it.