AI SEO

141 SEO Prompts You Can Run on Live SERP Data with Claude, ChatGPT, or Any AI Agent

Samuel Schmitt

This article gives you 141 SEO prompts for Claude, ChatGPT, Codex or any AI assistant.

With these SEO prompts, you can:

These prompts use live SERP data. Here’s how it works: you connect your AI assistant to thruuu, an SEO content optimization platform, through MCP tools.

thruuu is the bridge between your assistant and live SERP and AI Search data. It also gives your assistant thruuu’s other tools: keyword clustering, content briefs and the Content Pipeline.

Diagram of an AI assistant (Claude, ChatGPT, Codex) connected through MCP to thruuu, whose tools cover SERP analysis, search intent, AI Overview monitoring, brand mentions, topic clusters, content briefs, AI drafts and the Content Pipeline, and which reaches live Google SERP data and AI Search answers from ChatGPT, Gemini, Perplexity and Google AI Mode.

Before Starting, Connect thruuu’s SEO MCP Tools

You need a thruuu account, on any plan (yes, even on the free plan!), and the AI assistant you already use.

thruuu is an SEO content optimization platform: your assistant connects to its MCP server with your MCP key, and every prompt below then runs on your account.

There are two ways to connect.

The fastest way is to let your assistant do it. Paste this in, and it reads the instructions, asks for your MCP key and connects itself:

Set me up for thruuu using https://app.thruuu.com/agent-setup.md

You can also set it up yourself.

In thruuu, open API & Bulk Tools → MCP. It shows your MCP key and the exact setup for each assistant, key already filled in.

Then ask “What is my thruuu credit balance?” If your balance comes back, you’re connected.

Reading data is free. Before anything that uses credits, the assistant shows the cost and waits for your yes.

To learn more and understand all steps for each assistant, you can explore our SEO MCP server guide.

SEO Prompts for SERP Research and Competitor Analysis

Where does every content decision start? On the Google results page.

These prompts run a SERP analysis, read it back section by section, compare two keywords, and pull the full data behind each ranking page, down to its schema, headings, and outbound links.

Run an Analysis

Prompt 1: Run a SERP analysis

Run a SERP analysis for 'best project management software' in the US.

The assistant captures who ranks, People Also Ask, the AI Overview and more, usually in under a minute.

Prompt 1: Run a SERP analysis, shown in an AI assistant with the thruuu output

Prompt 2: Choose the market, device and depth

Analyse 'how to start a podcast' on Google UK, on mobile, first 2 pages, with search volume.

The same analysis with your market, device, depth (1 to 10 result pages) and search volume.

Prompt 2: Choose the market, device and depth, shown in an AI assistant with the thruuu output

Prompt 3: Analyse another country and language

Analyse 'cafetière pas cher' in France, in French, on google.fr.

The assistant picks the exact country, language and Google domain, so you never need a country code.

Prompt 3: Analyse another country and language, shown in an AI assistant with the thruuu output

Prompt 4: Analyse several keywords at once

Analyse these three keywords in the US: standing desk, ergonomic chair, desk lamp.

One analysis per keyword, all started at once and run side by side.

Prompt 4: Analyse several keywords at once, shown in an AI assistant with the thruuu output

Prompt 5: Save an analysis to a client workspace

Run a SERP analysis for 'crm for lawyers' in my Acme workspace.

The analysis is saved under that client or project.

Prompt 5: Save an analysis to a client workspace, shown in an AI assistant with the thruuu output

Find and Read Analyses

Prompt 6: List your recent analyses

Show my recent SERP analyses.

Your latest analyses, newest first, with date, AI engines included, whether a strategic report exists, and a link to each one in thruuu.

Prompt 6: List your recent analyses, shown in an AI assistant with the thruuu output

Prompt 7: Check before you pay twice

Do I already have an analysis for 'email marketing'?

The matching analyses, or a clear “not yet”.

Prompt 7: Check before you pay twice, shown in an AI assistant with the thruuu output

Prompt 8: List one client’s analyses

List the SERP analyses in my Acme workspace.

The same list, narrowed to one workspace.

Prompt 8: List one client's analyses, shown in an AI assistant with the thruuu output

Prompt 9: Read the top 10 results

Show me the top 10 results for my 'standing desk' analysis.

The top 10, including the AI Overview, People Also Ask and video slots, with volume, CPC and competition if you ran it with volume, and a link to the full analysis.

Prompt 9: Read the top 10 results, shown in an AI assistant with the thruuu output

Prompt 10: See the competitors’ outlines

Show me the outlines of the pages ranking for 'standing desk'.

Every ranking page’s H1 to H6 structure, side by side.

Prompt 10: See the competitors' outlines, shown in an AI assistant with the thruuu output

Prompt 11: Find the questions people ask

What questions do people ask around 'standing desk'?

People Also Ask plus the questions repeated in the ranking pages’ headings, ordered by how often they come up.

Prompt 11: Find the questions people ask, shown in an AI assistant with the thruuu output

Prompt 12: List the topics the top pages cover

What topics do the top pages for 'standing desk' cover?

The topics shared across the ranking pages, with how many pages use each one.

Prompt 12: List the topics the top pages cover, shown in an AI assistant with the thruuu output
What are Google's related searches for 'standing desk'?

The related searches Google shows for that keyword.

Prompt 13: Get the related searches, shown in an AI assistant with the thruuu output

Prompt 14: Read the search intent

What is the search intent behind 'standing desk'?

The intent read on the top results: a summary, the format split and the intent of each ranking page.

Prompt 14: Read the search intent, shown in an AI assistant with the thruuu output

Prompt 15: Find the heading and content patterns

What heading and content patterns do the top pages share?

The heading patterns and content patterns found across the ranking pages: the repeated structures and themes.

Prompt 15: Find the heading and content patterns, shown in an AI assistant with the thruuu output

Compare Analyses

Prompt 16: One article or two?

Should 'running shoes for flat feet' and 'best flat foot running shoes' be one article or two?

How many pages the two SERPs share, a similarity score, and a one-or-two recommendation using the same 40 percent rule as the app, with a link to the comparison.

Prompt 16: One article or two?, shown in an AI assistant with the thruuu output

Prompt 17: Find the pages that rank for both keywords

Which pages rank for both 'crm software' and 'best crm'?

The URLs both searches share, with both top 10s side by side.

Prompt 17: Find the pages that rank for both keywords, shown in an AI assistant with the thruuu output

Analysis Agent (the Strategic Report)

Prompt 18: Run the Analysis Agent

Run the Analysis Agent on my 'best crm' analysis.

A nine-section strategy report, ready in a few minutes, with a link to follow it.

Prompt 18: Run the Analysis Agent, shown in an AI assistant with the thruuu output

Prompt 19: List your strategic reports

Which of my keywords already have a strategic report?

Your reports with keyword, status (running, ready or failed) and date.

Prompt 19: List your strategic reports, shown in an AI assistant with the thruuu output

Prompt 20: Check a report’s status

Is my strategic report for 'best crm' ready?

Ready, still running, or failed with the reason and how to retry.

Prompt 20: Check a report's status, shown in an AI assistant with the thruuu output

Prompt 21: Read the full report

Give me the strategic report for 'best crm'.

All nine sections: summary, who is searching, intent and format, competitive landscape (top 8), AI engine landscape, brand landscape, topic coverage, recommendations and a content brief blueprint.

Prompt 21: Read the full report, shown in an AI assistant with the thruuu output

Prompt 22: Turn the report into a to-do list

From the 'best crm' report, what should my page include to win?

A short list of what to include, taken from the report’s recommendations and blueprint.

Prompt 22: Turn the report into a to-do list, shown in an AI assistant with the thruuu output

Run the Analyses from the Chat

Search Intent, Heading Analysis and Content Analysis start from the chat on any SERP analysis you already have.

They’re free, take a couple of minutes at most. You read the result by asking for it (Prompts 14 and 15).

Each prompt below has a longer version that asks for the full section in the same message.

Prompt 23: Run the search intent analysis

Run the search intent analysis on my 'standing desk' analysis.

The intent, the format table and the intent of every ranking page. If the analysis is linked to a brief, the brief’s Search Intent block fills in too, without overwriting anything you edited.

Run the search intent analysis on my "standing desk" analysis. When it has finished, show me the whole Search Intent section: the summary, the format table and the intent of every ranking page.
Prompt 23: Run the search intent analysis, shown in an AI assistant with the thruuu output

Prompt 24: Run the heading analysis

Run the heading analysis on my 'standing desk' analysis.

Heading clusters classed Very High, High, Medium or Low, plus the content opportunities.

Run the heading analysis on my "standing desk" analysis. When it has finished, show me the whole Heading Analysis section with the content opportunities.
Prompt 24: Run the heading analysis, shown in an AI assistant with the thruuu output

Prompt 25: Run the content analysis

Run the content analysis on my 'standing desk' analysis.

The repeated topics and content patterns with their classification, plus content opportunities.

Run the content analysis on my "standing desk" analysis. When it has finished, show me the whole Content Analysis section.
Prompt 25: Run the content analysis, shown in an AI assistant with the thruuu output

Read the Full Data of a SERP

Prompt 26: Get every page’s title, description, dates and schema

Show me the SERP title, description, dates and schema of every ranking page for 'standing desk'.

The Page Details table: Google’s title and description, the page’s meta description, page type, position, publication and modification dates, schema type, on-page FAQ count and link count.

Show me the SERP title, SERP description, meta description, page type, publication and modification dates and schema type of every result for my "standing desk" analysis. Give me the whole table, every row and every column, with no summary.
Prompt 26: Get every page's title, description, dates and schema, shown in an AI assistant with the thruuu output

Prompt 27: See every row of a results page

Show me everything on page 1 of my 'standing desk' analysis, with no 10-row limit.

Every row of the page in position order, including the AI Overview, People Also Ask and video slots. Ask for page 2, 3 and beyond on a multi-page analysis.

Show me every row of page 1 of my "standing desk" analysis, with no 10-row limit and no summary, together with the page details of each result. Then do the same for page 2 if the analysis has one.
Prompt 27: See every row of a results page, shown in an AI assistant with the thruuu output

Prompt 28: Read the SERP features in full

Show me the People Also Ask answers, video carousel, local pack and news for 'standing desk'.

People Also Ask with each answer and source, the video carousel, the local pack and the news items. A block that wasn’t on the SERP says so.

For my "standing desk" analysis, show me the People Also Ask questions with their answers, then the video carousel, the local pack and the news, in full.
Prompt 28: Read the SERP features in full, shown in an AI assistant with the thruuu output
Show me the FAQ and the links of the number 1 page for 'standing desk'.

The questions and answers on the page, then every outbound link with its label, URL, domain and rel.

For my "standing desk" analysis, show me the questions and answers in the FAQ of the result at position 1. Then list every link on that page with its label, URL, domain and rel.
Prompt 29: Read a page's FAQ and outbound links, shown in an AI assistant with the thruuu output

AI Search Prompts: ChatGPT, Gemini, Perplexity, Google AI Mode and AI Overviews

Ranking on Google and being cited in an AI answer are two different results. You need to see both.

The first group adds the AI engines to a SERP analysis, so you read what each one says and cites next to Google’s results.

The second builds AI Overview reports across hundreds of keywords and shows where your brand is cited, named, or missing.

AI Engines in a SERP Analysis

Prompt 30: Add ChatGPT to a SERP analysis

Run a SERP analysis for 'vegan protein powder' in the US and include ChatGPT.

Google’s results plus ChatGPT’s answer to the same query, in one analysis.

Prompt 30: Add ChatGPT to a SERP analysis, shown in an AI assistant with the thruuu output

Prompt 31: Compare four AI engines with Google

Analyse 'best crm' in the US with ChatGPT, Gemini, Perplexity and Google AI Mode.

One analysis holding Google plus the answers from ChatGPT, Gemini, Perplexity and Google AI Mode.

Prompt 31: Compare four AI engines with Google, shown in an AI assistant with the thruuu output

Prompt 32: Read what ChatGPT and Gemini answered

What did ChatGPT and Gemini say for 'standing desk', and which sources did they cite?

Each engine’s answer, its cited sources (top 10 each), and the brands and themes it named.

Prompt 32: Read what ChatGPT and Gemini answered, shown in an AI assistant with the thruuu output

Prompt 33: List every source one engine cited

Show every source Perplexity cited for 'standing desk'.

Perplexity’s full, uncapped source list.

Prompt 33: List every source one engine cited, shown in an AI assistant with the thruuu output

Prompt 34: See the outlines of the pages AI engines cite

What do the pages ChatGPT cites for 'standing desk' look like, heading by heading?

The outlines of the top 10 pages one engine cited.

Prompt 34: See the outlines of the pages AI engines cite, shown in an AI assistant with the thruuu output

Prompt 35: Find the brands the AI engines name

Which brands do the AI engines name for 'best crm', and what themes?

The brands each engine named and the themes tied to them.

Prompt 35: Find the brands the AI engines name, shown in an AI assistant with the thruuu output

Prompt 36: Read Google’s AI Overview and its sources

What does Google's AI Overview say for 'best crm', and who does it cite?

The AI Overview text, its cited sources, the brands and themes it names, and the questions found in those sources.

Prompt 36: Read Google's AI Overview and its sources, shown in an AI assistant with the thruuu output

Prompt 37: Add AI engines to a keyword you analysed

I only analysed Google. Add the AI engines to 'best crm'.

A fresh analysis with the engines you pick. The old one stays as it is.

Prompt 37: Add AI engines to a keyword you analysed, shown in an AI assistant with the thruuu output

AI Overview Reports

Prompt 38: Detect your brand name and aliases

What brand name and aliases would you track for example.com?

The brand name and aliases read from your home page, for you to confirm or edit. Nothing is created.

Prompt 38: Detect your brand name and aliases, shown in an AI assistant with the thruuu output

Prompt 39: Create an AI Overview report

Create an AI Overview report for example.com with these 120 keywords, US English, desktop.

The report started, with a link and the cost, once you confirm the domain, brand, aliases, keyword count and market. It never invents keywords or a market.

Prompt 39: Create an AI Overview report, shown in an AI assistant with the thruuu output

Prompt 40: Track one part of a site and several brand names

Create an AI Overview report for example.com/blog, tracking 'Acme' and its product lines.

The same report, scoped to a sub-path and tracking extra aliases or product lines.

Prompt 40: Track one part of a site and several brand names, shown in an AI assistant with the thruuu output

Prompt 41: Build a report from your own keyword sheet

Here is my keyword sheet with volumes. Check which keywords get an AI Overview.

The same report, built from the keywords you paste and using your volumes.

Prompt 41: Build a report from your own keyword sheet, shown in an AI assistant with the thruuu output

Prompt 42: List your AI Overview reports

List my AI Overview reports.

Your reports with label, domain, brand, type, status, whether it has data, keyword count, date and a link.

Prompt 42: List your AI Overview reports, shown in an AI assistant with the thruuu output

Prompt 43: Check whether a report has finished

Is my AI Overview report for example.com finished?

Whether at least one run has completed.

Prompt 43: Check whether a report has finished, shown in an AI assistant with the thruuu output

Prompt 44: Follow the visibility score over time

How has the visibility score of my main AI Overview report changed over its runs?

Every run with its date, status, keyword count, visibility score, domain citations, AI Overview presence and organic ranking.

Prompt 44: Follow the visibility score over time, shown in an AI assistant with the thruuu output

Prompt 45: Find where you win and where to act

Where am I winning in AI Overviews, and where is the biggest opportunity?

The latest run’s key metrics, opportunity views, priority actions, and top cited domains, mentioned brands and organic domains, with yours marked “(you)”.

Prompt 45: Find where you win and where to act, shown in an AI assistant with the thruuu output

Prompt 46: List the keywords where you’re cited and named

List the keywords where I am cited and named.

The winning view: keyword rows by search volume, up to 150.

Prompt 46: List the keywords where you're cited and named, shown in an AI assistant with the thruuu output

Prompt 47: Find mentions without a citation

Which keywords mention my brand but do not cite me?

The “capture a citation” view: keywords where the AI names you without linking to you.

Prompt 47: Find mentions without a citation, shown in an AI assistant with the thruuu output

Prompt 48: Find citations without a mention

Where am I cited or ranking but never named?

The “capture a brand mention” view: keywords where you’re a source but the answer never says your name.

Prompt 48: Find citations without a mention, shown in an AI assistant with the thruuu output

Prompt 49: See where one domain is cited

Which keywords cite youtube.com in their AI Overview?

The keywords citing that domain, the matching result blocks on the page (video, questions, forums) and whether the domain ranks top 10.

Prompt 49: See where one domain is cited, shown in an AI assistant with the thruuu output

Prompt 50: See what changed since the last run

What changed since the previous run?

New and lost mentions, new and lost citations (naming the competitor who took a lost one), new and lost top-10 rankings, AI Overviews that started or stopped appearing, and new branded keywords.

Prompt 50: See what changed since the last run, shown in an AI assistant with the thruuu output

Prompt 51: Compare two runs

Show me last month's run and compare it to the latest.

An earlier run’s numbers and what changed since.

Prompt 51: Compare two runs, shown in an AI assistant with the thruuu output

Prompt 52: Read one keyword’s full AI Overview

Show me the full AI Overview for 'best crm for small business' and its sources.

The complete AI Overview text, its full source list and the top 10 organic results, for one keyword in one run.

Prompt 52: Read one keyword's full AI Overview, shown in an AI assistant with the thruuu output

Prompt 53: Find the domains and brands Google cites most

Which domains and brands does Google cite most across my keywords?

The top cited domains, top mentioned brands and top organic domains across a run.

Prompt 53: Find the domains and brands Google cites most, shown in an AI assistant with the thruuu output

SEO Prompts for Keyword Clustering and Content Planning

How many articles does a 300-keyword list need? The list can’t tell you.

These prompts group keywords by the results they share, roughly one article per cluster. Then they show which clusters you already rank for, and which ones are gaps.

Create a Topic Cluster

Prompt 54: Group keywords into topic clusters

Group these 300 keywords into topic clusters for example.com, US English, top 20 results.

A link and a confirmation. Clustering runs in the background for several minutes, roughly one article per cluster.

Prompt 54: Group keywords into topic clusters, shown in an AI assistant with the thruuu output

Prompt 55: Compare clusters against your domain

Cluster this list, compare against my domain, and use a stricter overlap.

Clusters that show where you rank, with a stricter rule for how many shared results make two keywords belong together.

Prompt 55: Compare clusters against your domain, shown in an AI assistant with the thruuu output

Prompt 56: Cluster keywords with your own volumes

Cluster these keywords. I have volumes for them.

Clusters that carry your volumes, so you can sort by demand.

Prompt 56: Cluster keywords with your own volumes, shown in an AI assistant with the thruuu output

Prompt 57: Cluster a large keyword list

Cluster these 1,500 keywords for example.com, US English, top 20 results.

A link and a confirmation, as in Prompt 54. The assistant takes up to 2,000 keywords. For a longer list, it links you to the upload screen for an XLSX or CSV file.

Prompt 57: Cluster a large keyword list, shown in an AI assistant with the thruuu output

Read Your Topic Clusters

Prompt 58: List your topic clusters

List my topic clusters.

Your topic clusters, newest first, with label, date, number of clusters, status, target domain, results per keyword and a link.

Prompt 58: List your topic clusters, shown in an AI assistant with the thruuu output

Prompt 59: Check a topic cluster’s status

Is my 'Running shoes' topic cluster done?

Init, In Progress or Done.

Prompt 59: Check a topic cluster's status, shown in an AI assistant with the thruuu output

Prompt 60: See the clusters inside a topic cluster

Show the clusters inside my 'Running shoes' topic cluster.

Each cluster’s topic, its keywords (the first 15, then how many more), search volume, Avg PR, SERP features and brief status, plus your ranking and best page when a target domain is set.

Prompt 60: See the clusters inside a topic cluster, shown in an AI assistant with the thruuu output

Prompt 61: Find your content gaps

Which clusters am I not ranking for? Biggest volume first.

The clusters you don’t rank for, biggest volume first.

Prompt 61: Find your content gaps, shown in an AI assistant with the thruuu output

Prompt 62: Find the clusters you already rank for

Show the clusters I already rank for, with my best page for each.

The ranking clusters, with your best-ranking page and its title.

Prompt 62: Find the clusters you already rank for, shown in an AI assistant with the thruuu output

Prompt 63: See your favourite clusters

Which clusters have I marked as favourites?

The clusters you starred in the app, in their own column, treated as the ones you want to pursue.

Prompt 63: See your favourite clusters, shown in an AI assistant with the thruuu output

Prompt 64: Find the easiest gaps

Which clusters am I not ranking for that have the weakest competition?

The clusters you don’t rank for, with their Avg PR, lowest first. Avg PR is the average authority of the sites ranking for a cluster, the same number as on the cluster card in the app.

Prompt 64: Find the easiest gaps, shown in an AI assistant with the thruuu output

Prompt 65: Read a large topic cluster in full

List every cluster in my 'Email marketing' topic cluster, not just the first 150.

Every cluster, read page by page in the same order as Prompt 60, until the list ends.

Prompt 65: Read a large topic cluster in full, shown in an AI assistant with the thruuu output

Content Plan

Prompt 66: Build a content plan

Build a content plan for my 'Running shoes' topic cluster.

A create, optimise or skip table, with volume, SERP features, competition, competing domains, pages that rank for several clusters, totals and a link. When your domain matches a Brand Kit, its audience and positioning are used.

Prompt 66: Build a content plan, shown in an AI assistant with the thruuu output

Prompt 67: Filter the plan by intent and volume

Content plan, informational keywords above 500 searches only.

A follow-up to Prompt 66: only the matching clusters, with the full-topic totals unchanged.

Prompt 67: Filter the plan by intent and volume, shown in an AI assistant with the thruuu output

Prompt 68: Narrow the plan to one category

Only the 'Trail running' category.

Another follow-up to Prompt 66: the plan for one category of your topic cluster.

Prompt 68: Narrow the plan to one category, shown in an AI assistant with the thruuu output

Prompt 69: Find clusters by SERP feature

Which clusters trigger an AI Overview most often?

The clusters where a SERP feature (AI Overview, video, forums, featured snippet and others) shows at or above the level you set.

Prompt 69: Find clusters by SERP feature, shown in an AI assistant with the thruuu output

One Cluster

Prompt 70: See one cluster’s top results

Show the top results for the 'trail running shoes' cluster.

The top-ranking results merged across every search in the cluster.

Prompt 70: See one cluster's top results, shown in an AI assistant with the thruuu output

Prompt 71: Get a cluster’s outlines, questions and keywords

Show its outlines, common questions, intent and keywords.

A follow-up to Prompt 70: page outlines, People Also Asked and related searches, search intent, heading and content patterns (once run in the app), and up to 100 keywords.

Prompt 71: Get a cluster's outlines, questions and keywords, shown in an AI assistant with the thruuu output

Prompt 72: Analyse a cluster

Analyse the 'trail running shoes' cluster so I can see its results.

The ranking pages collected for a cluster that hasn’t been analysed, readable shortly after. It won’t run twice.

Prompt 72: Analyse a cluster, shown in an AI assistant with the thruuu output

Questions Across a Whole Topic

One call reads the People Also Ask questions and related searches of every cluster in a topic cluster, merges the duplicates and counts how often each one appears.

Prompt 73: Find the questions people ask across a topic

What questions do people ask across my 'Running shoes' topic cluster?

The 100 most frequent People Also Ask questions and related searches across every cluster, with how often each one appears and the clusters it comes from.

Prompt 73: Find the questions people ask across a topic, shown in an AI assistant with the thruuu output
Give me every People Also Ask question and related search in my 'Running shoes' topic cluster.

The full list, collected 300 at a time until nothing is left. A large topic cluster can hold 800 questions and 1,400 related searches.

Prompt 74: Get every question and related search, shown in an AI assistant with the thruuu output

Prompt 75: Find the questions that span the most clusters

Which questions come up in the most clusters of my 'Running shoes' topic cluster?

The questions shared by several clusters. They suit a pillar page or a hub FAQ rather than a single article.

Prompt 75: Find the questions that span the most clusters, shown in an AI assistant with the thruuu output

Prompt 76: Build an FAQ from the topic

Build an FAQ for my running shoes pillar page from the 20 most frequent questions in my 'Running shoes' topic cluster.

The 20 questions, grouped by theme, ready to add to a brief’s frequent questions (Prompt 95).

Prompt 76: Build an FAQ from the topic, shown in an AI assistant with the thruuu output
Which related searches in my 'Running shoes' topic cluster are not one of my keywords yet?

The related searches Google suggests that are missing from your keyword list.

Prompt 77: Find related searches you have not covered, shown in an AI assistant with the thruuu output

Prompts for Content Briefs and Outlines

A content brief is where research turns into instructions for a writer.

These prompts build one from a keyword, an analysis or a cluster. Then they read it back, edit any part of it, and rewrite the outline in a built-in style or one of your own.

Create a Brief

Prompt 78: Create a content brief

Create a content brief for 'how to start a podcast' in the US.

A brief with title, description and outline, built from a fresh SERP analysis in a minute or two. Find it afterwards with Prompt 83.

Prompt 78: Create a content brief, shown in an AI assistant with the thruuu output

Prompt 79: Use your own brief template

Make a brief for 'best standing desk' with my 'Product review' template.

A brief built on your saved template.

Prompt 79: Use your own brief template, shown in an AI assistant with the thruuu output

Prompt 80: Brief another market

Brief for 'standing desk' in the UK on mobile, with search volume.

The same brief, built on another market’s and device’s results.

Prompt 80: Brief another market, shown in an AI assistant with the thruuu output

Prompt 81: Turn an analysis into a brief

Turn my 'best crm' SERP analysis into a brief.

A brief built from the analysis you already have, with no new search. If one already exists for it, the assistant tells you which.

Prompt 81: Turn an analysis into a brief, shown in an AI assistant with the thruuu output

Prompt 82: Brief a cluster

Create a brief for the 'trail running shoes' cluster.

A brief built from the cluster’s existing data, with no new search. You give it a title.

Prompt 82: Brief a cluster, shown in an AI assistant with the thruuu output

Find and Read Your Briefs

Prompt 83: List your briefs

Show my briefs.

Your briefs, most recently updated first, with keyword, status, sections, last update and a link.

Prompt 83: List your briefs, shown in an AI assistant with the thruuu output

Prompt 84: Find a brief by topic

Find my briefs about podcasts.

The briefs matching a keyword or title.

Prompt 84: Find a brief by topic, shown in an AI assistant with the thruuu output

Prompt 85: List one client’s briefs

Which briefs are in my Acme workspace?

The briefs in one workspace.

Prompt 85: List one client's briefs, shown in an AI assistant with the thruuu output

Prompt 86: Read a whole brief

Show me everything in my 'standing desk' brief.

Every section: article summary, writer guidance, outline, search intent, competitor reference points, questions and the rest, with a link.

Prompt 86: Read a whole brief, shown in an AI assistant with the thruuu output

Prompt 87: Add the topics and competitor outlines

Include the topics and competitor outlines.

A follow-up to Prompt 86 that adds the two large sections normally left out: the top topics and the competing pages’ outlines.

Prompt 87: Add the topics and competitor outlines, shown in an AI assistant with the thruuu output

Prompt 88: Read just the outline

Show me just the outline of my 'standing desk' brief.

The heading structure only, with a link to the brief.

Prompt 88: Read just the outline, shown in an AI assistant with the thruuu output

Edit a Brief

Edits are free, but there’s no undo. The assistant reads the current value first and checks with you before saving. Editing needs Pay as you go or above.

Prompt 89: Change the tone and word count

Change the tone to friendly and expert, set the target to 1,800 words.

The article summary updated.

Prompt 89: Change the tone and word count, shown in an AI assistant with the thruuu output

Prompt 90: Set the audience and region

Set the audience to office managers in small businesses, in Canada.

The writer guidance updated: audience, region and notes.

Prompt 90: Set the audience and region, shown in an AI assistant with the thruuu output
Add these three links to the brief.

The brief’s links list replaced, keeping the ones you want.

Prompt 91: Add links to the brief, shown in an AI assistant with the thruuu output

Prompt 92: Add a note for the writer

Add a note: always mention our 30-day trial.

The text of a free text block replaced.

Prompt 92: Add a note for the writer, shown in an AI assistant with the thruuu output

Prompt 93: Rewrite the search intent summary

Rewrite the search intent summary in plain words.

The written summary in the search intent block replaced.

Prompt 93: Rewrite the search intent summary, shown in an AI assistant with the thruuu output

Prompt 94: Edit the topics

Add 'ergonomics' and 'sit-stand' to the topics, drop 'cheap'.

Topic terms added and removed.

Prompt 94: Edit the topics, shown in an AI assistant with the thruuu output

Prompt 95: Edit the frequent questions

Add a question about warranty and remove the one about price.

Frequent questions added and removed.

Prompt 95: Edit the frequent questions, shown in an AI assistant with the thruuu output

Edit the Outline

Prompt 96: Restructure the outline

Merge the second and third headings and add a section on common mistakes.

Your new outline, saved after your yes.

Prompt 96: Restructure the outline, shown in an AI assistant with the thruuu output

Prompt 97: Rewrite the outline in a style

Rewrite the outline of my 'standing desk' brief as a listicle.

A new outline in the chosen style, regenerated from the competitor pages in 30 to 60 seconds. The old one stays until the new one is ready.

Prompt 97: Rewrite the outline in a style, shown in an AI assistant with the thruuu output

Styles and Templates

Prompt 98: List the article types

What article types can I use?

Eight built-in types (tips, how-to, listicle, case study, product review, product alternative, comparison guide, ultimate guide) plus your own, with the instructions behind each.

Prompt 98: List the article types, shown in an AI assistant with the thruuu output

Prompt 99: Read how a style is written

Show me how the comparison guide style is written.

The structure rules behind one style.

Prompt 99: Read how a style is written, shown in an AI assistant with the thruuu output

Prompt 100: Save your own outline style

Save a new outline style called 'Buyer's guide': intro with a verdict, comparison table, pros and cons per product, FAQ.

A reusable style, selectable in the app and by the assistant.

Prompt 100: Save your own outline style, shown in an AI assistant with the thruuu output

Prompt 101: Edit your own style

Make my 'Buyer's guide' style more creative.

Your style updated: name, instructions or creativity. Only the fields you mention change.

Prompt 101: Edit your own style, shown in an AI assistant with the thruuu output

Prompt 102: List your brief templates

Which brief templates can I use?

Your saved templates and the built-in default.

Prompt 102: List your brief templates, shown in an AI assistant with the thruuu output

Prompts for Writing AI Drafts and Running a Content Pipeline

This is where a brief becomes an article.

The first group writes and checks one AI draft at a time. The second sends batches of keywords through the whole chain, from analysis to draft, and stops once for your review.

AI Drafts

Prompt 103: Write a full draft

Write a full draft from my 'how to start a podcast' brief.

A draft started from your brief, with a link. It takes a few minutes, and there’s one draft per brief.

Prompt 103: Write a full draft, shown in an AI assistant with the thruuu output

Prompt 104: List your drafts

Show my drafts.

Your drafts, newest first, with title, status, word count, to-do count, date and link.

Prompt 104: List your drafts, shown in an AI assistant with the thruuu output

Prompt 105: Find the draft with the most work left

Which draft has the most work left?

Your drafts sorted by open to-dos, starting with the one that has the most.

Prompt 105: Find the draft with the most work left, shown in an AI assistant with the thruuu output

Prompt 106: Read a draft

Read the draft for 'how to start a podcast'.

The full article, its word count and a link to the draft in thruuu.

Prompt 106: Read a draft, shown in an AI assistant with the thruuu output

Prompt 107: List what’s left to fill in

What do I still need to fill in on that draft?

A follow-up to Prompt 106: a checklist of stats to source, claims to verify, links to find and calls to action to write, each under its heading.

Prompt 107: List what's left to fill in, shown in an AI assistant with the thruuu output
Which links could not be placed in the draft?

The unplaced links, with the reason for each.

Prompt 108: Find the links the draft couldn't place, shown in an AI assistant with the thruuu output

Prompt 109: Check a draft against your Brand Kit

Read my draft and check it against our Brand Kit.

Where the draft follows or drifts from your tone, rules and words to avoid, with suggested rewrites.

Prompt 109: Check a draft against your Brand Kit, shown in an AI assistant with the thruuu output

Content Pipeline

The Content Pipeline takes up to 20 keywords at a time through SERP analysis, deeper analysis, brief and draft, with one review stop before the draft. Professional and above.

Prompt 110: Set up a Content Pipeline

Set up a Content Pipeline for my workspace.

Your pipeline, ready for keywords, after a one-time setup the assistant asks about first.

Prompt 110: Set up a Content Pipeline, shown in an AI assistant with the thruuu output

Prompt 111: Send a batch of keywords through the pipeline

Create content for these 10 keywords in the US with my Brand Kit.

All ten keywords added and launched once you confirm the market, template, article type, AI engines, Brand Kit and total cost, with a link to the Content Pipeline.

Prompt 111: Send a batch of keywords through the pipeline, shown in an AI assistant with the thruuu output

Prompt 112: Choose the style, template and market

Add these keywords, how-to style, no Brand Kit, UK English.

The same batch with your article type, template and market, and the Brand Kit switched off.

Prompt 112: Choose the style, template and market, shown in an AI assistant with the thruuu output

Prompt 113: Add AI engines to a batch

Run these keywords through the pipeline and include ChatGPT.

AI engine answers collected for every keyword in the batch.

Prompt 113: Add AI engines to a batch, shown in an AI assistant with the thruuu output

Prompt 114: Check the pipeline’s progress

How are my pipeline keywords doing?

Each keyword’s status, any error and what to do next, with a link to the Content Pipeline.

Prompt 114: Check the pipeline's progress, shown in an AI assistant with the thruuu output

Prompt 115: See what’s waiting for your review

Which keywords are waiting for my review?

The items at the review stop, whose brief is ready to read.

Prompt 115: See what's waiting for your review, shown in an AI assistant with the thruuu output

Prompt 116: Read a brief before you approve

Show me the brief for 'best crm for startups'.

The brief and the analysis behind it.

Prompt 116: Read a brief before you approve, shown in an AI assistant with the thruuu output

Prompt 117: Edit the brief at the review stop

Tighten that outline and set the audience, then I will approve.

A follow-up to Prompt 116: the brief edited, with your OK on each change.

Prompt 117: Edit the brief at the review stop, shown in an AI assistant with the thruuu output

Prompt 118: Approve a brief and start the draft

Approve the brief for 'best crm for startups' and start the draft.

The brief approved and the draft generating, once you confirm the cost. It takes a few minutes.

Prompt 118: Approve a brief and start the draft, shown in an AI assistant with the thruuu output

Prompt 119: Approve every waiting brief

Approve every brief that is awaiting review.

Several drafts started, after the assistant lists the items and states the total.

Prompt 119: Approve every waiting brief, shown in an AI assistant with the thruuu output

Prompt 120: Read a finished pipeline draft

Show me the finished draft for 'best crm for startups'.

The full article and its to-do list, the same as Prompt 106, reached from the pipeline.

Prompt 120: Read a finished pipeline draft, shown in an AI assistant with the thruuu output

Prompt 121: Archive finished items

Archive the items whose drafts are ready.

Finished items archived to free their slots. The brief and draft stay readable.

Prompt 121: Archive finished items, shown in an AI assistant with the thruuu output

Prompt 122: Check your free pipeline slots

How many pipeline slots do I have left?

Slots used out of 20.

Prompt 122: Check your free pipeline slots, shown in an AI assistant with the thruuu output

Prompt 123: Find failed or stuck keywords

Which keywords failed or ran out of credits?

The stuck items, each with the reason and what to do next.

Prompt 123: Find failed or stuck keywords, shown in an AI assistant with the thruuu output

Multi-Step SEO Prompts: Whole Workflows in One Message

One message, several steps.

The assistant confirms the cost before each spend, and each example shows the final answer with one line per step it took. The step lists point back to the single prompts above, so you can also run any content workflow one step at a time.

Prompt 124: Keyword to finished draft

Run a SERP analysis for 'best crm for startups' in the US with ChatGPT included. Show me what the top pages cover and what ChatGPT says. Then create a brief from it, rewrite the outline as a comparison guide, and once I approve, write the draft.

Runs Prompts 1 and 30, 12, 32, 81, 97 and 103, and returns an analysis, a brief, a restyled outline and a draft.

Prompt 124: Keyword to finished draft, shown in an AI assistant with the thruuu output

Prompt 125: Keywords to pipeline, review and drafts

Create content for these 10 keywords in the US with my Brand Kit and my Product review template. Tell me the total cost first. When the briefs are ready, show me which need review.

Runs Prompt 110 if needed, then 111, 115, 116, and 118 or 119, and returns ten briefs to review, then drafts.

Prompt 125: Keywords to pipeline, review and drafts, shown in an AI assistant with the thruuu output

Prompt 126: Find gaps, then fill them

Look at my 'Running shoes' topic cluster for example.com. Show the clusters where I do not rank, biggest volume first. Take the top 5 keywords, send them through the Content Pipeline in the US with my Brand Kit, and tell me the total cost before you start.

Runs Prompts 61, 66 and 111, and returns a gap list, then five briefs on the way.

Prompt 126: Find gaps, then fill them, shown in an AI assistant with the thruuu output

Prompt 127: AI Overview report to page fixes

Create an AI Overview report for example.com with these 80 keywords. When it is ready, show me where I am winning and where I could capture a citation. For the top three opportunities, show the full AI Overview and suggest what my page should add.

Runs Prompts 38, 39, 45, 47 and 52, then Prompts 89 to 95, and returns a report, priorities and concrete content changes.

Prompt 127: AI Overview report to page fixes, shown in an AI assistant with the thruuu output

Prompt 128: Analysis Agent to brief update

Run the Analysis Agent on my 'best crm' analysis. Summarise the competitive landscape and which brands the AI engines mention. Then update my brief's audience and topics to match the recommendations.

Runs Prompts 18, 21, 35, 90 and 94, and returns a strategy report and a brief aligned to it.

Prompt 128: Analysis Agent to brief update, shown in an AI assistant with the thruuu output

Prompt 129: One article or two, then brief

Compare my analyses for 'crm for real estate' and 'real estate crm software'. If the SERPs overlap, build one brief for both. If not, build two.

Runs Prompt 16, then Prompt 81 once or twice, and returns a clear decision and the matching briefs.

Prompt 129: One article or two, then brief, shown in an AI assistant with the thruuu output

Prompt 130: Plan a month of content

Build a content plan for my 'Email marketing' topic cluster: informational keywords above 300 searches, split into create and optimise. Pick 10 to create, add them to the pipeline in a how-to style, and list them for my review when the briefs are ready.

Runs Prompts 66, 67, 112 and 115, and returns a plan, ten keywords in motion and a review list.

Prompt 130: Plan a month of content, shown in an AI assistant with the thruuu output

Prompt 131: See what AI engines say, then brief it

Show me what ChatGPT and Perplexity say for 'vegan protein powder' and which sources they cite. If my domain is missing, create a brief from this analysis that fixes that.

Runs Prompts 32, 33 and 81, and returns the engine answers, their cited sources and a brief that targets the gap.

Prompt 131: See what AI engines say, then brief it, shown in an AI assistant with the thruuu output

Prompt 132: Review drafts against the Brand Kit

Read the Brand Kit for Acme, then check my three newest drafts against it. List what to rewrite in each.

Runs Prompts 141, 104 and 109, and returns a punch list per draft, in the client’s voice.

Prompt 132: Review drafts against the Brand Kit, shown in an AI assistant with the thruuu output

Prompt 133: Weekly agency check across clients

For each of my workspaces, show how the pipeline is doing, what is waiting for my review, and what changed in the latest AI Overview report.

Runs Prompts 137, 114, 115 and 50, and returns a roundup per client.

Prompt 133: Weekly agency check across clients, shown in an AI assistant with the thruuu output

Setup and Credit Prompts

Want to know what you can run before you spend anything? Start here.

Your balance, your plan, your workspaces, the markets you can search and your Brand Kits. All free to ask, on any plan unless noted.

Prompt 134: Check your credit balance

How many credits do I have left?

The credits available. On a company account, that’s the shared company balance, the same number the app header shows.

Prompt 134: Check your credit balance, shown in an AI assistant with the thruuu output

Prompt 135: Check your plan and role

What plan am I on, and am I the account owner?

Your plan, whether you’re the account owner, and your balance.

Prompt 135: Check your plan and role, shown in an AI assistant with the thruuu output

Prompt 136: Check you can afford a batch

Can I afford to analyse 20 keywords and write briefs for all of them?

A plain yes or no, with the estimated cost against your balance.

Prompt 136: Check you can afford a batch, shown in an AI assistant with the thruuu output

Prompt 137: List your workspaces

Which workspaces do I have?

Each workspace your account belongs to, with the default one marked. Many accounts have none, and that’s fine.

Prompt 137: List your workspaces, shown in an AI assistant with the thruuu output
Which countries, languages and Google domains can I search in?

The countries, languages and Google domains thruuu supports, plus the device and results-depth options.

Prompt 138: List the markets you can search, shown in an AI assistant with the thruuu output

Prompt 139: Check which AI engines your plan includes

Which AI engines can I include on my plan?

Which of ChatGPT, Gemini, Perplexity and Google AI Mode your plan allows, and which plan unlocks the others.

Prompt 139: Check which AI engines your plan includes, shown in an AI assistant with the thruuu output

Prompt 140: List your Brand Kits

Show me my Brand Kits.

Every Brand Kit you can access, with brand name, workspace (or Personal) and last refresh date.

Prompt 140: List your Brand Kits, shown in an AI assistant with the thruuu output

Prompt 141: Read a full Brand Kit

Show me the full Brand Kit for my agency client Acme.

One kit in full: identity, audience, value proposition, proof points, tone, writing rules, words to avoid, style guide and call-to-action style.

Prompt 141: Read a full Brand Kit, shown in an AI assistant with the thruuu output

How to Write SEO Prompts That Work

Every prompt above is plain language. Rewrite any of them in your own words.

What separates a prompt that returns the data from one that returns a tidy paragraph about it? These nine habits.

Say the Job, the Market and the Output You Want

A prompt that works every time has three parts:

  1. Say the job in one sentence: “Show me the search intent for my analysis of ‘standing desk’.”
  2. Name the market in plain words: “United States, English, desktop”. The assistant looks up the exact values. A market it gets wrong is refused, so be specific.
  3. Describe the shape of the answer: “the full table, every row and column, no summary”, or “one section only”.

Do you need a spending rule too? No.

Showing the cost and waiting for your yes is already the default, so it protects you even when you forget.

Feed It Real SERP Data, Not a Guess

An assistant with no data still answers.

Ask it for keyword difficulty, search volume or who ranks, and you get plausible numbers.

Through thruuu, the data comes from a live search: the ranking pages, their outlines, People Also Ask, the AI Overview and the AI engine answers.

So make the analysis step one. “Analyse ‘standing desk’ in the US, then tell me what the top pages cover” gets you an answer built on the pages Google ranks today. Drop the first half and you get the model’s memory of the topic.

Give It Your Context

The same keyword means different work for different sites. Tell the assistant whose content this is.

Say “in the Acme workspace” and the analysis, brief or report lands with that client. Name a Brand Kit and the assistant reads your audience, positioning, tone and words to avoid before it plans a cluster (Prompt 66), builds a pipeline batch (Prompt 111) or reviews a draft (Prompt 109). Add a target domain to a topic cluster and it knows where you already rank.

Give it a keyword, not an ID. The assistant finds the right analysis, workspace, brief or report for you. If you do paste an ID, copy it exactly.

Ask for a Format, and Ask for the Raw Data

By default, the assistant summarises what it reads.

Want the data itself? Say so: “the whole table, every row and every column”. Or add “Show me the raw data” to the end:

Show me the results of my "standing desk" analysis with the SERP title, meta description, dates and schema of every result. Show me the raw data.
Run the search intent, heading and content analyses on my "standing desk" analysis. When they have finished, show me the raw data.

For big data, ask one question per message: the results table, then the outlines, then the intent. Shorter answers are less likely to be condensed.

Make It Show Its Work

Ask for the original wording: “Quote exactly, keep the original language, tell me if you shortened anything.” The AI Overview and the engine answers come back in full, and the assistant is asked to flag any shortening.

Ask which analysis, engine and run an answer came from, too.

Most answers end with an “Open in thruuu” link to the analysis, brief, report or draft behind them. When a number matters, click it. The app shows the same data the assistant read.

Verify the Numbers Before You Use Them

Check a figure against the linked analysis before it goes into a report or a brief.

Comparing with an export from the web app? Match the settings. Device, number of pages, language and volume all change the results.

Watch the totals. Lists show a slice and state the true total: the list of analyses shows the 20 most recent (50 at most), and other lists say “…and N more” rather than hiding rows. Narrow by keyword or workspace to find an older item.

Separate Research from Writing

Run the research prompts first (SERP research and AI Search), and read what they return before you ask for a brief or a draft.

Why? A brief built on an analysis you’ve read is a brief you can correct. A brief built in the same breath as the research isn’t.

The Content Pipeline builds this in. It runs the analysis and the brief, then stops at a review step (Prompt 115). Nothing is drafted until you approve.

Reuse Before You Re-Run, and Keep an Eye on Credits

Before you analyse a keyword, ask “Do I already have an analysis for this keyword?” (Prompt 7). It’s free, and an analysis you already paid for can become a brief with no new search (Prompt 81).

Reading, listing, comparing and checking never cost anything. Starting new work does, and the assistant states the cost and waits for your yes every time.

Check your balance first (“How many credits do I have left?”) and ask Prompt 136 before a big batch. Every cost is in the table in the next section.

Treat the Output as a Starting Point, Not a Final Answer

A brief is a plan. A draft is a draft.

thruuu doesn’t invent facts in a draft. It leaves a marker where a stat, a claim, a link or a call to action needs a human, and Prompt 107 lists them. Prompt 109 checks the draft against your Brand Kit.

The finishing happens in the thruuu app or your own editor. The assistant can’t edit a draft.

What SEO Prompts Can’t Do: Limits and Costs

What’s free? Reading, listing, comparing and checking. Always.

Anything that starts new work costs credits, and the assistant states the cost and waits for your yes first.

If you don’t have enough, nothing is charged and you’re told how short you are. On a company account, credits are pooled.

Here’s what each job costs:

What you ask forApproximate cost in credits
SERP analysis1, plus 1 for volume, plus 1 per AI engine
Brief5, plus 1 for volume
Brief from an existing analysis or cluster5
Analysis Agent5
AI draft15
Topic cluster or AI Overview reportKeywords divided by 20, 10 or 5 depending on depth, rounded up
Analyse a cluster1
Content Pipeline6 per keyword for analysis (plus engines and volume), 5 for the brief, 15 for the draft after you approve
Search intent, heading and content analysesFree

The connection works on every plan, and anything your plan doesn’t include replies with a short upgrade note and charges nothing.

New analyses, briefs, clusters, reports and drafts take from under a minute to a few minutes. The assistant starts them and tells you where to look.

Edits to briefs and outlines are free but have no undo, so the assistant checks with you first.

And here’s what the connection doesn’t do:

SEO Prompts FAQ

Can ChatGPT or Claude Generate SEO Content?

Yes, of course they can. The content gets better when you connect Claude or ChatGPT to thruuu’s Content Pipeline, because the research happens before the writing.

The pipeline analyses the SERP, collects what AI search engines answer for the keyword, and compiles everything into a brief with a structured outline. Then it writes the draft from that brief. The Content Pipeline needs the Professional plan or above, and the AI engines are optional.

We cover the full setup in our guide to writing SEO content with Claude.

Can I Rely on ChatGPT or Claude for Keyword Research?

Yes, but there are limits. Without a data source, the assistant estimates search volume and rankings from memory, and the numbers look real.

Connected to thruuu, it runs a live search for any keyword and shows you the pages that rank. Ask for search volume and you also get the monthly volume and the CPC, for 1 extra credit.

How Do I Verify an AI’s SEO Suggestions?

You’re the SEO expert. The assistant is there to support your process.

Be aware that it will always try to give you an answer. So bring your own knowledge, and challenge the ideas and suggestions your AI assistant gives you.

Run These 141 Prompts with Your AI Assistant

Connect thruuu to Claude, ChatGPT, Codex or the AI agent you already use, and every prompt in this list runs on live SERP data.