Skip to content

Playbook · Marketing and e-commerce · Amazon sellers and e-commerce founders

Find a product worth selling on Amazon. With demand, rivals and a better angle.

Amazon product research checks whether a product idea has demand, who already wins it, and what you could do better. With Agentik, your AI measures Amazon search demand and reads the listings that win it through DataForSEO, checks who advertises the product on Meta with Adspy.sh, and finds what buyers complain about on Reddit with Exa.

Get $5 free credit See the prompt

providers
3
tools
7
example, at most
$0.18

Prices from the live catalog · updated

One request to your AI3 providers, 6 steps, one connection
  1. Step 1: Measure Amazon demandDataForSEO$0.03
  2. Step 2: See which listings winDataForSEO$0.06
  3. Step 3: Check the seasonDataForSEO$0.02
  4. Step 4: See who advertisesAdspy.sh$0.04
  5. Step 5: Hear what buyers complain aboutExa$0.06
  6. Step 6: Score the opportunityYour AI, no tool call

Paid from one prepaid balanceat most $0.18

Why 3 providers

Each provider does one part. Your AI does the rest.

No single provider covers this job. Through Agentik your AI calls each one for its part, from one connection.

  1. DataForSEO

    Amazon searches and their volume, the listings that win with their prices and sellers, and Google Trends seasonality.

    5 tools · steps 1, 2 and 3$0.09

  2. Adspy.sh

    Brands advertising the product on Meta, and the angles they keep running.

    1 tool · step 4$0.04

  3. Exa

    Reddit threads where buyers say what existing products get wrong.

    1 tool · step 5$0.06

Without Agentik

3 accounts, 3 API keys, 3 bills.

  • DataForSEOSign-up, card, API key, bill
  • Adspy.shSign-up, card, API key, bill
  • ExaSign-up, card, API key, bill

Then code to call each API, retry it and join the results before your AI can use them.

With Agentik

One connection. One prepaid balance.

  • Amazon and Google data, Meta ads and Reddit search come from three providers through one connection.
  • The product search runs as a task at DataForSEO’s lower price; your agent quotes everything before running.
  • Your AI shows the price before each call, and a call that ends in an error is refunded.

Step by step

How to research a product to sell on Amazon in Claude or ChatGPT. Step by step.

Prices are for this example: One idea on amazon.com, 100 related searches, 5 listings with their sellers, 5 terms over 5 years, 20 Meta ads, 3 Reddit searches. They are the most each call can reserve at today’s catalog prices.

  1. Step 1: Measure Amazon demand

    DataForSEO

    Related Amazon searches with their monthly volume.

    Tools called in step 1
    Amazon related keywordsdataforseo/dataforseo-labs-amazon-related-keywords100 searchesat most $0.03
  2. Step 2: See which listings win

    DataForSEO

    The products that rank, then the top five in detail: price, rating, reviews and every seller.

    Tools called in step 2
    Amazon product searchdataforseo/amazon-products-task1 searchat most $0.0015
    Amazon product page by ASINdataforseo/amazon-asin5 listingsat most $0.03
    Amazon sellers and offers for an ASINdataforseo/amazon-sellers5 listingsat most $0.03
  3. Step 3: Check the season

    DataForSEO

    Google Trends over five years shows whether demand is steady, growing or seasonal.

    Tools called in step 3
    Google Trendsdataforseo/google-trends5 terms, 5 yearsat most $0.02
  4. Step 4: See who advertises

    Adspy.sh

    Brands running Meta ads for the product, sorted by how long each ad has run.

    Tools called in step 4
    Search ad creativesadspy/search-creatives20 adsat most $0.04
  5. Step 5: Hear what buyers complain about

    Exa

    Exa searches Reddit by meaning for problems people have with the products they bought.

    Tools called in step 5
    Web searchexa/search3 searchesat most $0.06
  6. Step 6: Score the opportunity

    Your AI

    Your AI weighs demand, competition and the complaints into a go or no-go.

What you get

What you get: a go or no-go with the evidence. In your conversation.

You provide

  • A product idea
  • The marketplace (amazon.com by default)

Your AI returns

  • Amazon searches for the idea, with monthly volume
  • The top listings: price, rating, reviews and sellers
  • Five years of seasonality, and the brands advertising on Meta
  • The complaints a better product could fix, with a go or no-go
Opportunity scoreExample layout
Product ideaMonthly searchesTop priceSellersComplaint to fix
Oak standing deskWobbles at full height
Compact deskCable mess
Desk with drawersSlow motor

Placeholders: your rows come from your run.

Give this to your agent
Use my connected Agentik MCP server for this job: amazon product research. Research a product idea for Amazon. First ask me for the idea and the marketplace. Find related Amazon searches and their volume with dataforseo/dataforseo-labs-amazon-related-keywords. List the products that rank with dataforseo/amazon-products-task, then read the top five with dataforseo/amazon-asin and their sellers with dataforseo/amazon-sellers. Check five years of seasonality with dataforseo/google-trends. Find brands advertising the product on Meta with adspy/search-creatives. Search Reddit for what buyers complain about with exa/search. Finish with demand, competition, seasonality, the complaints a better product could fix, and a go or no-go with your reasons. Before running anything, inspect each tool with inspect_tool, ask me for any input you don’t have instead of inventing it, and tell me the total price. Wait for my go. At the end, tell me what Agentik charged.

The prompt

Paste it into Claude, ChatGPT, Codex or Cursor.

Once Agentik is connected, your AI asks for your inputs, checks each tool and shows the price before anything runs.

Or ask it your way

  • “Is there room for a premium bamboo cutting board on amazon.com?”
  • “Which standing desks win on Amazon, at what price, and who sells them?”
  • “Is demand for heated vests seasonal, and when does it start?”
  • “What do Reddit users dislike about current pet water fountains?”

What it costs

At most $0.18. From one balance.

Each provider’s own price, with 0% markup. Each call reserves at most the amount shown, is billed for what it actually used, and the rest is released. A call that ends in an error is refunded.

This exampleat today’s catalog prices

DataForSEO$0.09

  • Amazon related keywords100 searches$0.03
  • Amazon product search1 search$0.0015
  • Amazon product page by ASIN5 listings$0.03
  • Amazon sellers and offers for an ASIN5 listings$0.03
  • Google Trends5 terms, 5 years$0.02

Adspy.sh$0.04

  • Search ad creatives20 ads$0.04

Exa$0.06

  • Web search3 searches$0.06

Total, at most$0.18

Good to know

  • Search volumes are DataForSEO estimates for Amazon, good for comparing ideas.
  • Sales figures aren’t public; reviews, rank and the number of sellers are the signals used.

When to use something else

To track sales estimates and keyword ranks every day, an Amazon seller suite fits better. Agentik suits researching ideas with your AI.

Questions

Amazon product research: straight answers.

Still unsure? Ask us.

Which Amazon marketplaces work?

The marketplaces DataForSEO supports; your agent checks yours before running.

Couldn’t I just use DataForSEO directly?

For the Amazon and Trends data, yes. Who advertises on Meta and what buyers complain about on Reddit come from Adspy.sh and Exa. Through Agentik your AI uses all three from one connection; the example costs at most $0.18.

What does the example cost?

At most $0.18.

Keep going

More playbooks for your AI.

  • Firecrawl, Adspy.sh and fal.ai

    Product photo ads

    Your product page for the copy, your category’s longest-running ads as the brief, then new scenes, headlines and a clip.

    3 providers7 toolsexample at most $1.09

  • Exa, Firecrawl, DataForSEO, Hacker News and X

    Market research

    The companies in the space, their pricing pages, Google search demand for the category, and what buyers say on Hacker News and X.

    5 providers7 toolsexample at most $0.43

  • Adspy.sh, Firecrawl and Internet Archive

    Competitor ad research

    Their longest-running Meta and Google ads, why they work, where they send people, and how those offers changed.

    3 providers6 toolsexample at most $0.67

Browse the tools: E-commerce · SEO · Advertising

Run it today

Start with $5 on us. Then just ask.

Sign in with Google or GitHub, connect your AI with one link, then paste the prompt. Eligible new accounts get $5 of credit for low-cost calls.

Get $5 free credit Connection guides

No card · No subscription · 0% markup on tool calls