Perplexity's experimental sandbox for deep research agents, data analysis, and long-form reports that used to take hours.
Perplexity Labs is the experimental tier that sits above standard Perplexity Search. Where vanilla Perplexity answers one question well, Labs chains multiple searches, file uploads, and code execution into a full research task — spreadsheet-crunching, multi-source synthesis, and long-form reports that Google-search alone can't match.
Who it's for: Analysts, researchers, and consultants who run 5+ deep-dive queries per week. If you mostly need one-off quick answers, standard Pro search is enough.
Labs decomposes a big question ("compare Q3 churn drivers across 20 SaaS peers") into sub-queries, runs each, cites sources, and assembles a structured report. Comparable to OpenAI's Deep Research but tuned for spreadsheet output.
Ask for a competitor matrix or pricing teardown and Labs renders a real table, not markdown. Export to CSV or copy-paste into Google Sheets. This is the killer use case that hooked us.
Drop PDFs, CSVs, or screenshots and Labs answers grounded questions on them. Combined with web search, it synthesizes internal docs + public sources in one reply.
Runs Python in the sandbox for quantitative questions. Plot trends over time from raw CSV data, compute descriptive stats, or transform files without leaving the chat.
Perplexity Labs is the strongest "agentic research" option for analysts who live in spreadsheets. It's not as flashy as OpenAI Deep Research, but for pricing teardowns, competitive matrices, and CSV-grounded analysis it's the best return on $20/month in 2026.