420 AI Prompts for Research

The best Research prompts across 6 subcategories. Tested on ChatGPT, Claude, Gemini and every major model.

Browse the prompts

You are a comedy analyst specializing in extracting and cataloging humor from any text—transcripts, articles, social media, scripts, or conversations.

# Task

Identify every distinct joke in the content below. For each one, capture both the setup and the punchline in a two-line bullet format that preserves the comedic timing and structure.

# Rules

- Extract **only** jokes; ignore commentary, analysis, or non-humorous content
- Format each joke as a bullet with the setup on the first line and the punchline indented on the second
- Skip duplicates or minor retellings of the same joke
- Preserve the original wording and style (casual, observational, one-liner, callback, etc.)
- Output clean Markdown with no preamble, warnings, or notes

# Example structure

- Why don't scientists trust atoms?  
  Because they make up everything.

- I told my wife she was drawing her eyebrows too high.  

Joke Extractor From Text

Generate a summary of jokes extracted from any given text content, producing a Markdown bulleted list with setups and punchlines.

4
You are an expert at extracting counterintuitive insights from written content—the ideas that challenge assumptions, reveal hidden patterns, or reframe familiar concepts in unexpected ways.

Your specialty lies in **existential, cognitive, and technological themes**: the purpose and meaning of life, human flourishing, AI's impact on humanity, the evolution of learning and culture, memes, books, continuous improvement, and the future shaped by technology.

# TASK

Analyze the content below and extract **10 genuinely surprising or novel insights**—not summaries, but the non-obvious takeaways a casual reader would miss.

For each insight:
1. Identify the **counterintuitive claim or pattern** (not the obvious thesis)
2. Capture it in **exactly 8 words** using Paul Graham's voice: conversational, concrete, free of jargon and academic phrasing
3. Rank by **surprise × importance**—prioritize id

Powerful Insights Extractor

Generate concise, counterintuitive insights from various content sources in Paul Graham's voice.

4

Video Commerce Entity Extractor

Extracts commercially relevant entities from video transcripts and categorizes them for analysis. Produces an actionable list of products, services, and brands based on viewer engagement potential.

4
You are a research assistant specializing in bibliographic extraction. Your task is to identify and list every mention of creative works, academic sources, and learning materials from the provided content.

# Scope
Extract references to:
- Books (fiction, non-fiction, poetry collections)
- Academic papers, journal articles, and research publications
- Films, plays, television series, and other dramatic works
- Visual art (paintings, sculptures, installations) and exhibitions
- Music albums, compositions, and performances
- Essays, articles, and think pieces
- Historical documents and primary sources
- Podcasts, lectures, and educational media

# Instructions
1. Read the content carefully and identify all explicit and implicit references to the above categories.
2. For each reference, capture the title and creator/author when available.
3. Preserve the original spelling and formatting of 

References And Citations Extractor

Generates a list of references and citations from provided content, ideal for research assistance.

4
You are an expert analyst skilled at distilling complex content into its most valuable essence—the single insight that matters most.

# YOUR TASK

Analyze the provided content and identify the one idea that is most important, surprising, or actionable. Then crystallize it into two precise statements: the core insight itself, and what someone should do with it.

# ANALYSIS APPROACH

1. Read the entire {{content}} thoroughly to understand context, arguments, and conclusions
2. Identify competing ideas and rank them by significance, novelty, and practical impact
3. Select the single most valuable takeaway—prioritize insights that challenge assumptions or enable decisions
4. Formulate the idea in its purest form, free of jargon or hedging
5. Derive a concrete recommendation that follows logically from the insight

# OUTPUT FORMAT

Return exactly two sections in Markdown:

## MAIN IDEA
[Exact

Main Idea Extractor

Generate a structured summary by extracting the main idea and recommendation from complex content.

4

MCP Server Mentions Extractor

Generates a thorough analysis of MCP server mentions from provided content, identifying specific features, integrations, and usage patterns.

4
# ROLE

You are a specialist in extracting and structuring activity timelines from unstructured transcripts, logs, and recordings. Your strength lies in identifying discrete events, recognizing contextual relationships between them, and presenting a clear chronological narrative.

# TASK

Analyze the provided transcript or log and extract all significant activities, events, and interactions. Identify temporal sequences, causal links, and thematic connections between events to construct a coherent activity map.

# INPUT

{{transcript}}

# ANALYSIS STEPS

1. **Parse the content**: Read through the entire transcript to establish overall context, participants, setting, and timeframe
2. **Identify discrete events**: Mark each distinct activity (conversations, media consumption, gameplay, transitions, decisions, interactions)
3. **Map relationships**: Note which events share context, participa

Main Activities Extractor From Transcript

Generate a 16-word summary and a list of key events from a transcript, highlighting main activities.

3
You are a research analyst trained in critical evaluation of claims against established scientific consensus. Your task is to identify and extract statements from the provided text that contradict widely accepted scientific evidence or established facts.

# Scope of Extraordinary Claims

Extract statements that:
- Contradict conclusions supported by overwhelming scientific consensus (e.g., evolution, climate change, vaccine efficacy, germ theory)
- Promote conspiracy theories debunked by credible evidence (e.g., moon landing denial, flat Earth, chemtrails)
- Assert supernatural causation for natural phenomena without basis
- Misrepresent or deny historical events with strong evidentiary support

# Analysis Process

1. Read the entire input carefully to understand context and intent
2. Identify statements that meet the criteria above—consider both explicit claims and implied positions
3. 

Extraordinary Claims Extractor

Extracts extraordinary claims from a conversation by identifying statements that contradict established scientific consensus.

2

What are AI prompts for Research?

AI prompts for Research are engineered instructions that already work. These are not one-line questions. Each one fixes the role, the context, the task and the output format before you type a word, so you get a usable result on the first run instead of the fourth.

They cover the work Research actually get asked for: research and briefs, copy and content, analysis and reporting, planning, outreach and the admin that eats the day. Open a card to see the full prompt and the output it returns.

The deepest areas here: Market Research (116 prompts), Competitive Analysis (59 prompts), Trend Analysis (50 prompts), Literature Review (47 prompts), Deep Research (40 prompts).

Popular on this page right now: "Joke Extractor From Text", "Powerful Insights Extractor", "Recommendations Extractor From Content".

420 on this page, every one scoped to Research. Free to read, free to copy.

Why these prompts work for Research

A weak prompt costs you the hour you were trying to save: you rewrite it three times, get something generic, then finish the job by hand. An engineered prompt front-loads that thinking once.

In Research that means first drafts you can send, analysis you can act on, and the repetitive work handed off, so the time goes into judgement instead of typing.

Every prompt here was written for a real job and tested against the models people actually use. Nothing scraped from a thread.

How to use these prompts

Open a prompt, copy it, and replace the [bracketed] variables with your own product, audience or topic. The structure around them stays as is. That structure is the part doing the work.

Paste it into ChatGPT, Claude, Gemini, Grok or the model you already use. If the output drifts, tighten the context line instead of rewriting the whole prompt.

No account needed to copy one. No setup, no extension, nothing to install.

Which AI tool works best for Research prompts?

Text prompts here run well in ChatGPT, Claude, Gemini and Grok; image prompts target Midjourney and Nano Banana. Each card lists the models it was tested with.

Are these AI prompts free to use?

A big part of the library is free: open a prompt, copy it, use it. Premium packs and the Complete AI Bundle unlock the full collection with lifetime updates.

How do I adapt these prompts to my use case?

Start with the [variables]: niche, audience, constraints. If the result still misses, add one example of the output you want. A single good example beats three extra instructions.

For a prompt built from scratch, the Start Now card above opens the custom prompt generator.

Related resources

Get smarter on AI every week

One email a week with the best new prompts, tools, and model updates. Unsubscribe anytime.

Join 100,000+ subscribers. One email a week, real prompts, tools, and model updates. Unsubscribe anytime.

Keyword research prompts

A model cannot see search volumes, so a keyword research prompt should ask for what it can do: expand a seed into the questions people ask, group them by intent and name the page type each group needs. Take that list into your keyword tool for the numbers. The prompts in this category follow that order and skip the fake volume estimates.

Seed keyword: [keyword]. Audience: [who]. List 40 related queries people type into Google, grouped into: informational, comparison, transactional, problem. For each group, name the page type that would rank (guide, list, tool page, product page) and the one query to target first. No volume numbers.

Research paper prompts

Research paper prompts do the slow parts: turning a question into a searchable set of terms, summarising a paper into claims and evidence, and drafting a literature review paragraph from your notes. They never invent citations, because the prompt tells them not to.

Here are my notes on five papers: [paste]. Write a literature review paragraph for each theme you find, citing only the papers in my notes by author and year. Where two papers disagree, say so in one sentence. Flag any claim in my notes that has no source attached.