Are Perplexity and GPT Researcher both free?

GPT Researcher is entirely free and open-source (Apache 2.0) — you only pay for the LLM and search-API keys you supply yourself. Perplexity has a free tier with limited Pro searches and a few Deep Research runs a day, then $20/month for Pro (~20 Deep Research runs/day), $200/month for Max (unlimited), and Enterprise from $40/seat/month.

Which one can research my own private documents, not just the web?

GPT Researcher — point it at local files (PDF, Word, CSV, Markdown and more) and it researches and cites across them, alongside or instead of the open web. Perplexity's research is scoped to the live web (plus, on its Enterprise tier, connectors into business systems like Snowflake and Salesforce) — no Perplexity tier lets you upload and research arbitrary private documents the way GPT Researcher does.

Do I need to install anything to use either one?

GPT Researcher, yes — it's a developer tool you run yourself as a Python package, a FastAPI + Next.js app, or in Docker, supplying your own LLM and search-API keys. Perplexity needs no installation at all: use it on the web, in its mobile apps, or through its Comet browser.

Which came first, GPT Researcher or Perplexity's Deep Research?

GPT Researcher. Assaf Elovic open-sourced it in 2023, and its README calls it "the first open deep research agent" (its own claim about open-source implementations, not an independently audited fact). Perplexity didn't ship its own Deep Research feature until February 14, 2025 — GPT Researcher's plan-search-synthesize pattern predates it by well over a year.

Can either be used programmatically, via an API or MCP?

Both. Perplexity offers the usage-based Sonar API plus an official MCP server. GPT Researcher can be run as a FastAPI service you call yourself, and a companion project (gptr-mcp, not bundled in the main repo) adds an MCP server on top of it.

Perplexity vs GPT Researcher

Perplexity's Deep Research is a polished, hosted subscription answer engine — ask a question, get a cited report in minutes, no setup — while GPT Researcher is a free, open-source, self-hosted research-agent framework you install and run against your own LLM and search-API keys, so the real choice is a done-for-you research subscription versus an open, ownable research-automation engine.

Feature	Perplexity Visit Perplexity ↗	GPT Researcher Visit GPT Researcher ↗
Description	The AI answer engine, with agentic Deep Research and the Comet browser that acts on the web for you.	An open-source autonomous research agent that plans, searches many sources and writes a cited report.
Category	Research & data agents	Research & data agents
Pricing	Free	Free
Tier	Listed	Listed
API	Yes	Yes
Tags	Browser useEnterpriseClaudeAPIGPTAutonomousFree tier	Self-hostedMulti-agentAPIOpen sourceOpen modelsAutonomousFree tier

Perplexity and GPT Researcher both call their flagship research feature "Deep Research," and both sit in this index's research-agents category, but they take opposite approaches. Perplexity is a hosted, closed-source AI answer engine: ask it a question and its Deep Research mode autonomously runs dozens of searches, reads the results and writes a structured, cited report in a few minutes, wrapped in a polished web, mobile and desktop-browser (Comet) product with a flat, tiered subscription. GPT Researcher is the opposite bet — an Apache 2.0-licensed, open-source Python framework you install and run yourself: a planning agent breaks your question into sub-questions, executor agents fan out parallel web searches, and a publisher agent assembles the accumulated context into a long-form, cited report, with no hosted product wrapped around it at all.

The shared name is not a coincidence of two teams landing on the same phrase independently — GPT Researcher genuinely came first. Its creator, Assaf Elovic, open-sourced it in 2023, and its own README describes it as "the first open deep research agent designed for both web and local research" — a claim about being first among open-source implementations, not an independently audited fact, but the timeline checks out: Perplexity didn't launch its own Deep Research feature until February 14, 2025, over a year later. So the two aren't rivals that emerged at the same moment — GPT Researcher popularized the plan-then-parallel-search-then-synthesize architecture as an open reference implementation well before most buyers had heard the phrase "deep research," and Perplexity (along with OpenAI's and Google's own versions) later turned that same pattern into a polished, hosted product feature.

The practical differences follow from that split. Perplexity is far simpler to start with — no installation, just sign in and ask a question — with Deep Research included from the $20/month Pro tier (roughly 20 runs a day) and unlimited runs on the $200/month Max tier, which also unlocks Perplexity's broader agentic Computer tools. GPT Researcher carries no subscription at all: it's free forever, and you supply your own LLM and search-API keys, with its own documentation citing a typical Deep Research run at around five minutes and well under a dollar in API costs — cheaper per run at low volume, in exchange for the setup work (installing the package, wiring up API keys, picking a runtime) Perplexity doesn't require. The two also research different territory: Perplexity is scoped to the live web, plus, on Enterprise, connectors into business systems like Snowflake and Salesforce; GPT Researcher can instead (or additionally) point at your own local documents — PDFs, Word files, CSVs — and cite across them, a capability no Perplexity tier offers.

Verdict

Want a polished, no-setup research assistant you can use in a browser or app right now, with per-query model choice across Sonar, GPT, Claude and Gemini and an agentic browser included → Perplexity ($20/month Pro or $200/month Max). Want a free, open, self-hosted research engine you can embed in your own product, point at whichever model and search API you choose, and use to research private documents as well as the web → GPT Researcher. Running many reports a day and want to do the math first → GPT Researcher's pay-per-run costs (cents to about a dollar per run) can undercut a flat subscription at low-to-moderate volume, while Perplexity's flat fee wins once usage is high enough that per-run costs would clear $20-200 a month anyway. Need an enterprise product with SSO, business-system connectors and vendor support → Perplexity Enterprise; GPT Researcher has no enterprise tier or vendor support at all — you self-support whatever you deploy.

Visit Perplexity ↗Visit GPT Researcher ↗

The digest

Decided? Watch the bench move. New agents, fresh verdicts, moves in the rankings, one short email, every Tuesday.

Frequently asked questions

Are Perplexity and GPT Researcher both free?: GPT Researcher is entirely free and open-source (Apache 2.0) — you only pay for the LLM and search-API keys you supply yourself. Perplexity has a free tier with limited Pro searches and a few Deep Research runs a day, then $20/month for Pro (~20 Deep Research runs/day), $200/month for Max (unlimited), and Enterprise from $40/seat/month.
Which one can research my own private documents, not just the web?: GPT Researcher — point it at local files (PDF, Word, CSV, Markdown and more) and it researches and cites across them, alongside or instead of the open web. Perplexity's research is scoped to the live web (plus, on its Enterprise tier, connectors into business systems like Snowflake and Salesforce) — no Perplexity tier lets you upload and research arbitrary private documents the way GPT Researcher does.
Do I need to install anything to use either one?: GPT Researcher, yes — it's a developer tool you run yourself as a Python package, a FastAPI + Next.js app, or in Docker, supplying your own LLM and search-API keys. Perplexity needs no installation at all: use it on the web, in its mobile apps, or through its Comet browser.
Which came first, GPT Researcher or Perplexity's Deep Research?: GPT Researcher. Assaf Elovic open-sourced it in 2023, and its README calls it "the first open deep research agent" (its own claim about open-source implementations, not an independently audited fact). Perplexity didn't ship its own Deep Research feature until February 14, 2025 — GPT Researcher's plan-search-synthesize pattern predates it by well over a year.
Can either be used programmatically, via an API or MCP?: Both. Perplexity offers the usage-based Sonar API plus an official MCP server. GPT Researcher can be run as a FastAPI service you call yourself, and a companion project (gptr-mcp, not bundled in the main repo) adds an MCP server on top of it.

Still deciding? Get a quote from Perplexity

Tell us what you need. We make the introduction to Perplexity.