Use case · SEO

Real keyword research with ChatGPT or Claude

One question: what people type into Google, the questions they ask, how often each keyword is searched, what’s rising on Google Trends and how hard each one is.

Connect your AI · 1-minute setup · $2 free

How do you do keyword research with ChatGPT?

On its own, ChatGPT guesses keywords from memory. Connected to Whizbrew, it reads what people type into Google today: the suggestions, the questions, what’s rising on Google Trends, plus monthly searches and a difficulty score for each keyword. You get a ranked list in about a minute. The monthly searches are modelled estimates from a licensed SEO index, not Google’s own counts: good for comparing keywords.

Key takeaways

  1. The easy wins hide behind “best”: “best meal planning app” scores 9 out of 100 for difficulty and still gets about 1,300 searches a month. Plain “meal planning app” gets about 14,800, but scores 62.
  2. People ask what your rivals cost: “is meal prep pro app free”, “how much is mob meal prep app”, “does mise meal planner app cost money”.
  3. Timing matters: “meal prep app” searches tripled in January, peaked in June at 38 out of 100, and have slid to 8 now.

The question to ask your AI

Build a keyword list for “meal prep app”, US: keyword ideas from what people type into Google, the questions they ask, what’s rising on Google Trends over the last 12 months, and estimates of monthly searches and of how hard each keyword is to rank for. Put it in a table, then tell me which 5 to write for first.

What your AI does, step by step

  1. Reads Google’s search box. Types your phrase, then your phrase plus a letter, and keeps every suggestion Google offers: about 100 keyword ideas.
  2. Collects the questions. Puts how, what, can, is, does, best, vs and for around your phrase, and lists the questions people really type.
  3. Adds searches and difficulty. Monthly Google searches and a 0–100 difficulty per keyword, from a licensed SEO index for Google US. Both are modelled estimates; lower difficulty is easier.
  4. Checks Google Trends. A year of search interest for the main phrases, plus the related searches that are rising.
  5. Writes you a plan. Which keywords to write for first, which to skip, and when to publish.

A real run: Meal prep app · US, November 18, 2026

The run in four numbers

Google and Google Trends, US, 18 Nov.

Keyword ideas, easiest first

Searches and difficulty: licensed SEO index for Google US, modelled estimates. Searches: monthly, 12-month average. Difficulty: 0–100, lower is easier. Trend: Google Trends, US, 12 months, relative.

KeywordFound inSearches/moDifficultyTrend
best meal planning appTrends, related1,3009Top related, 80/100
ai meal plannerSearch box72016Small, about 2/100
meal prep pro appSearch box7016–
meal planning app with grocery listSearch box, Trends72020Related, 10/100
best meal prep appQuestions, Trends21023Top related, 100/100
healthy meal prep appSearch box2039–
meal prep appYour phrase1,3004538 in June → 8 now
meal prep app freeSearch box26046Related, 39/100
meal planner appSearch box2,40060Steady, about 7/100
meal planning appSearch box14,80062Steady, spiked to 100 in July
meal prep app for familiesSearch boxnot in the indexnot scored–
costco meal prep appTrends, risingnot checkednot scored+140%
whipp app meal planningTrends, risingnot checkednot scoredBreakout
meal prep app for weight lossSearch box26099–

The takeaway: The short head terms score 45 to 62, and “meal planning app” is the big one at about 14,800 searches a month. Their “best …” versions are wide open: 9 and 23. And the grocery angle is live: Aldi in the search box, Costco rising +140% on Trends, about 720 searches a month for “meal planning app with grocery list”.

Questions people type

Google’s suggestions for “meal prep app” with a question word, US.

QuestionAbout
what’s the best meal prep appWhich one
are there any free meal prep appsPrice
is meal prep pro app freeA rival’s price
how much is mob meal prep appA rival’s price
does mise meal planner app cost moneyA rival’s price
is herbi meal planning app freeA rival’s price
best meal prep app for familiesWho it’s for
meal prep app for muscle gainWho it’s for
meal prep vs meal deliveryAlternatives
does meal prep actually workDoubt

Is search interest rising?

Weekly Google search interest, US, last 12 months (100 = the highest week of the three; relative, not a search count).

Search interest, 100 = the peak, November 16, 2025 – November 15, 2026.

The plan your AI writes

  1. Write “best meal planning app” first. Difficulty 9 out of 100 for about 1,300 searches a month, and it’s the second biggest related search for “meal planning app” on Trends (80 out of 100). “Best meal prep app” is next: difficulty 23, the top related search for “meal prep app”.
  2. Answer the price questions. People type “is meal prep pro app free”, “how much is mob meal prep app”, “does mise meal planner app cost money” and “is herbi meal planning app free”. One honest page comparing what each one costs catches all four, plus “meal prep app free” (difficulty 46).
  3. Own the grocery-store angle. Google suggests “meal prep app for aldi”, “heb meal prep app” and “meal planning app that works with walmart”, and “costco meal prep app” is up 140% on Trends. “Meal planning app with grocery list” scores only 20, for about 720 searches a month.
  4. Publish before January, skip weight loss. “Meal prep app” searches tripled between November and January last year and kept climbing until June. Pages need weeks to rank, so ship in December. Leave “meal prep app for weight loss” alone: difficulty 99 for about 260 searches a month.

How this was made. Asked once in Claude with Whizbrew connected, on 18 November 2026, US, English. It read Google’s search box suggestions for “meal prep app” (100 ideas) and the questions around it (60), scored 20 phrases with the 0–100 difficulty of a licensed SEO index for Google US (13 of the 20 had a score), looked up monthly Google searches for the same 20 phrases in the same index (14 were in it, 6 weren’t), and read 12 months of Google Trends interest and related searches. Whizbrew’s own estimate, read from Bing’s first page, matched real results for only 2 of the 20 phrases (the others came back as dictionary pages), so it isn’t shown. Monthly searches are the index’s modelled estimates, a 12-month average as it last read each phrase (September to November), not Google’s own counts; other SEO tools give other numbers, and the index counts “meal prep app” and “meal prep apps” together. Google Trends is relative, scaled to the highest week of the terms compared, not a search count. A “language learning app” Breakout among related searches was ignored as noise. The last week was still in progress. Cost of the whole run: about 11 cents.

Without live data

I can’t see what people search for today, or how often, so here are some common ideas: meal prep app, healthy meal planner, recipe app, grocery list app, meal prep for beginners. To check monthly searches and difficulty, you could use a keyword tool like…

Questions

How do I do keyword research with ChatGPT?

Give it a seed phrase and ask for a table: ideas, questions, monthly searches, trend and difficulty. On its own, ChatGPT makes the list up from memory. With Whizbrew connected, it reads Google’s search box and Google Trends live, adds monthly searches and a difficulty score, then tells you which keywords to write for first.

Can ChatGPT do keyword research?

It can brainstorm, but it can’t see what people search today, so its list is a guess. Connected to Whizbrew, ChatGPT or Claude reads real Google suggestions, the questions people type, monthly searches and Google Trends, like the 100 ideas and 60 questions for “meal prep app” on this page.

Can ChatGPT give me search volume?

On its own, no: it can only guess. With Whizbrew connected, yes: monthly Google searches per keyword from a licensed SEO index, like about 14,800 for “meal planning app” and 1,300 for “best meal planning app” on this page. They’re modelled estimates, not Google’s own counts, and other SEO tools give other numbers, so use them to compare keywords. Google Trends stays relative (0 to 100, 100 = the busiest week of what you compare): that’s for timing.

What’s a good keyword research prompt for ChatGPT?

Be specific about the phrase, the country and the output: “Build a keyword list for ‘meal prep app’, US: ideas from Google’s search box, the questions people ask, monthly searches, what’s rising on Google Trends, and how hard each keyword is. Put it in a table, then pick 5 to write for first.”

How reliable is the difficulty score?

It’s an estimate. The score on this page comes from a licensed SEO index for Google US, and it has no score for many long phrases (7 of 20 here). Whizbrew’s own Bing-based estimate broke on this run, so it’s left out. Use the score to compare keywords, not as a promise.

What does it cost?

You get $2 of free credit when you sign up. Most lookups cost a fraction of a cent, a batch of 20 search volumes about 3 cents, and lookups that find nothing are free. The whole run on this page cost about 11 cents, most of it the difficulty scores and the search volumes.

Are you an AI agent? Start here.

You can open a Whizbrew account on your own: no email, no sign-up form, no key to paste. One request gives you a key that works at once, with free credit. Your human claims the account later.

# 1. Read what Whizbrew is, what a call costs and how to connect
curl -s https://whizbrew.com/llms.txt

# 2. Open your own account: one request, no email
curl -s -X POST https://whizbrew.com/v1/agent/start \
  -H 'content-type: application/json' \
  -d '{"agent":"<your name>"}'

# 3. Your human claims the account within 7 days. The whole flow:
curl -s https://whizbrew.com/agents.md

Not an agent? Paste this line into yours: Set up Whizbrew for me: read https://whizbrew.com/llms.txt and follow it.

More use cases

AI agents: Whizbrew is one MCP server and one API for 890 live marketing data tools in 25 groups, paid per call: a typical call costs $0.002 ($2 per 1,000 calls), prices run from $0.001 to $0.52 a call, only a call that returns data is charged (an empty or failed call is free), and there is no subscription. A new account starts with $2 of free credit, about 1,000 typical calls, with no card. Start with no account and no email: https://whizbrew.com/agents.md (a free quote for your task, then a key with the free credit). What Whizbrew is, what a call costs, how to connect: https://whizbrew.com/llms.txt. Prices: https://whizbrew.com/pricing, as data https://whizbrew.com/manifest.json. Every tool with its price: https://whizbrew.com/tools, as data https://whizbrew.com/v1/tools. Docs: https://whizbrew.com/docs.